Article Text

Download PDFPDF

Effect of HIV-1 and antiretroviral therapy on herpes simplex virus type 2: a prospective study in African women
  1. P Mayaud1,
  2. N Nagot1,3,4,
  3. I Konaté3,
  4. A Ouedraogo3,
  5. H A Weiss2,
  6. V Foulongne4,
  7. M-C Defer3,
  8. A Sawadogo5,
  9. M Segondy4,
  10. P Van de Perre4,
  11. on behalf of the ANRS 1285 Study Group
  1. 1
    Clinical Research Unit, Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, UK
  2. 2
    Infectious Diseases Epidemiology Unit, Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK
  3. 3
    Centre Muraz, Bobo-Dioulasso, Burkina Faso
  4. 4
    Université Montpellier 1, EA 4205 “Transmission, Pathogenèse et Prévention de l’Infection par le VIH”, and Laboratoire de Bactériologie-Virologie and Département d’Information Médicale, CHU Montpellier, Montpellier, France
  5. 5
    University Hospital of Bobo-Dioulasso, Bobo-Dioulasso, Burkina Faso
  1. Dr P Mayaud, Clinical Research Unit, Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London WC1E 7HT, UK; philippe.mayaud{at}lshtm.ac.uk

Abstract

Objectives: To document the natural history of herpes simplex virus type 2 (HSV-2) in relation to HIV and highly active antiretroviral therapy (HAART) in Africa, a longitudinal study was conducted of women in the placebo arms of two randomised controlled trials of HSV-suppressive therapy in Burkina Faso.

Methods: 22 HIV-uninfected women (group 1), 30 HIV-1-infected women taking HAART (group 2), and 68 HIV-1-infected women not eligible for HAART (group 3) were followed over 24 weeks. HSV-2 DNA was detected on alternate weeks using real-time PCR from cervicovaginal lavages. Plasma HIV-1 RNA was measured every month. CD4 cell counts were measured at enrolment.

Results: Ulcers occurred on 1.9%, 3.1% and 7.2% of visits in groups 1, 2 and 3 (p = 0.02). Cervicovaginal HSV-2 DNA was detected in 45.5%, 63.3% and 67.6% of women (p = 0.11), and on 4.3%, 9.7% and 15.5% of visits in the three groups (p<0.001). Among HIV-infected women, cervicovaginal HSV-2 DNA was detected more frequently during ulcer episodes (adjusted risk ratio (aRR) 2.79, 95% CI 2.01 to 3.86) and less frequently among women practising vaginal douching (aRR 0.60, 95% CI 0.40 to 0.91). Compared with women not taking HAART and with CD4 cell counts of 500 cells/μl or greater, women on HAART had a similar risk of HSV-2 shedding (aRR 0.95, 95% CI 0.52 to 1.73), whereas women with CD4 cell counts of 200–500 cells/μl were more likely to shed HSV-2 (aRR 1.71, 95% CI 1.02 to 2.86).

Conclusions: HSV-2 reactivations occur more frequently among HIV-infected women, particularly those with low CD4 cell counts, and are only partly reduced by HAART. HSV therapy may benefit HIV-infected individuals during HAART.

View Full Text

Statistics from Altmetric.com

Herpes simplex virus type 2 (HSV-2) is one of the most common sexually transmitted infections (STI) worldwide.1 2 HSV-2 has synergistic bidirectional relationships with HIV: HSV-2 may enhance the risk of HIV acquisition3 and the risk of HIV transmission among dually infected individuals through increased HIV genital shedding.47 HSV-2 may influence HIV disease progression because clinical and subclinical HSV reactivations trigger increases in plasma HIV-1-RNA levels,8 9 which, over time, might precipitate a more rapid decline of the CD4 cell count.10 Furthermore, HIV alters the natural history of HSV-2 infection. Some studies in north American populations have shown that patients co-infected with HIV and HSV-2 may experience more severe and prolonged genital ulcerations, particularly if profoundly immunosuppressed.11 HIV-infected patients also present with more frequent genital HSV-2 recurrences,12 13 including more frequent asymptomatic reactivations,4 5 12 which are partly driven by immunosuppression. Immune recovery induced by highly active antiretroviral therapy (HAART) has been associated with fewer and less severe HSV-2 clinical recurrences, although HSV-2 shedding persisted even when plasma HIV-1 RNA was undetectable and CD4 cell counts had increased.14 15 On the other hand, severe recurrences of herpesvirus infections, including mucocutaneous herpes, have been associated with immune reconstitution syndrome following HAART initiation, probably associated with very low baseline CD4 cell counts.16 17

HSV-2 has become the leading cause of genital ulcer disease (GUD) in many sub-Saharan African countries1820 but little is known of the epidemiology and natural history of HSV-2 in Africa. HSV-2 seroprevalence in the general population ranges from 24% in West Africa to 74% in South Africa, mirroring the gradient of HIV infection rates.1 Cross-sectional studies have assessed HSV-2 genital shedding in relation to HIV-1,4 5 but the longitudinal pattern of clinical or subclinical herpes reactivations has not been described, in particular among HIV-infected individuals and those using HAART.

We recently conducted two randomised controlled trials of HSV-suppressive treatment among HSV-2 and HIV-1 co-infected women taking HAART21 or not eligible for HAART,6 21 in which we reported the impact of 500 mg twice daily valaciclovir on plasma and genital HIV-1 RNA and on genital HSV-2 DNA. In the present study, we compared the occurrence of GUD and HSV-2 genital shedding among women enrolled in the placebo groups of these trials, and in a group of HSV-2-seropositive and HIV-uninfected women also taking placebo as a “masking group” in the trials. The present study provides the first longitudinal data on the natural history of HSV-2 infection in HIV-infected and uninfected populations in Africa.

PATIENTS AND METHODS

Study design and participants

Dually HIV-1 and HSV-2-seropositive women from Bobo-Dioulasso, Burkina Faso, with no immediate indication for HSV-2-suppressive therapy (less than six GUD episodes in the previous 12 months) were enrolled in two separate trials according to antiretroviral therapy requirement, as described.6 21 As part of the screening process for trial enrolment, we also identified HSV-2-seropositive women without HIV infection. We included a random sample of these women as a “masking group” to preserve the confidentiality about HIV serostatus in this closely knit community of high-risk women. Except for HIV serostatus, the eligibility criteria were identical to those of women enrolled in the trials. HIV-uninfected women were given the same placebo, as none of them was eligible for HSV-suppressive therapy. Both the local investigators and the participants were blinded to treatment allocation status.

The research protocol was approved by the institutional review board at Centre Muraz and by the research ethics committees at the Burkina Faso Ministry of Health and at the London School of Hygiene and Tropical Medicine. The trials were registered with the International Clinical Trials Registers (NCT00158509).

Study procedures

Eligibility was assessed at a screening visit during which venous blood samples were collected to detect serum antibodies for HIV-1, HSV-2 and syphilis, to assess renal and liver functions and to obtain CD4 cell counts, as described.6 21 Women requiring HAART were first treated with antiretroviral drugs for a period of at least 4 months before trial enrolment. Eligible women who provided witnessed written consent were enrolled into a baseline phase (without study drug) with six visits on alternate weeks over a 12-week period, before random assignment to study drugs and a further follow-up period of 12 weeks with six visits on alternate weeks. Full details of the trial procedures and HAART regimen have been reported previously.6 21 The present analysis includes only women who were allocated to placebo, including the group of HIV-negative women.

At each of the visits, a doctor examined the patients, diagnosed GUD and collected genital specimens, including swabs for the diagnosis of vaginal infections and a cervicovaginal lavage enriched with cervical swabbing, which was aliquoted (with no centrifugation step) and stored at −80°C. One aliquot was used to detect and quantify HSV-2 DNA using real-time PCR (lower limit of detection 500 copies/ml), as described.22

Menstruating women were deferred for genital sampling until 2 days after bleeding had ceased. At the first visit, cervical swabs for the diagnosis of Neisseria gonorrhoeae and Chlamydia trachomatis by PCR were also obtained. Plasma HIV-1 RNA was quantitated monthly in HIV-infected women using real-time PCR with quantitation thresholds of 300 copies/ml,23 and the CD4 cell count was determined (FACScan; Becton Dickinson, Franklin Lakes, New Jersey, USA) at enrolment.

All participants were counselled on HSV-2 infection at enrolment and requested to return promptly if they experienced any GUD episode between follow-up visits. Lesional swabs were collected from clinically confirmed ulcers (but not vesicles) before any treatment was initiated, and tested for HSV-2 DNA by type-specific PCR using the same methodology and thresholds as for cervicovaginal lavage enriched with cervical swabbing.6 21 22 We did not test for any other GUD aetiology. HIV counselling, free STI medication and supplies of condoms were offered at each visit. GUD treatment followed the World Health Organization modelled national guidelines, with syndromic antibiotic first-line treatment, followed 7 days later, for non-healing ulcers, with aciclovir 200 mg five times a day for 5 days.

Statistical methods

We defined herpetic ulcers as GUD cases presenting as typical vesicles or those in which lesional HSV-2 DNA was detected. The frequencies of GUD and cervicovaginal HSV-2 DNA occurring at least once were compared across groups using Poisson regression with robust standard errors. Further analyses were based on visits using repeated-measures analyses. The frequencies of GUD and HSV-2 shedding were compared between groups with Poisson regression using generalised estimating equations with an exchangeable correlation matrix. The quantity of genital HSV-2 DNA (log10 copies/ml) among visits with detectable HSV-2 DNA was analysed using random effects linear regression.

We assessed the determinants of HSV-2 genital shedding among HIV-infected women using a Poisson regression model with robust standard errors. Variables with p<0.20 on univariable analysis, or variables previously shown to be associated with HSV-2 shedding (eg, bacterial vaginosis),24 were included in the final multivariable model. HIV status was categorised in three mutually exclusive subgroups: (1) women taking HAART; (2) women not taking HAART with CD4 cell counts of 200–500 cells/μl and (3) women not taking HAART with CD4 cell counts of 500 cells/μl or more.

All analyses were conducted using Stata version 9.0.

RESULTS

Groups 1, 2 and 3 were composed of 22 HIV-uninfected women, 30 HIV-1-infected women taking HAART and 68 HIV-1-infected women not eligible for HAART, respectively.

The characteristics of study participants are presented in table 1. All women were HSV-2 seropositive and none were taking any antiherpetic drugs. The mean age was comparable between groups (33.9, 35.0 and 32.6 years, respectively). Sex workers represented 100%, 30% and 70% of women in groups 1, 2 and 3, respectively. For women in group 2, the median time since HAART initiation was 20.3 weeks (interquartile range 17.6–26.4). Although the majority of women in group 2 (28/30) had undetectable plasma HIV-1 RNA levels, they still had a lower median CD4 cell count compared with HIV-infected women in group 3. Approximately 90% of women in all groups practised vaginal washing at least once a day. STI and other genital infections were rare in all three groups, with the exception of bacterial vaginosis, which was much more prevalent among HIV-infected women (groups 2 and 3).

Table 1 Characteristics of study participants at enrolment

GUD episodes and HSV-2 cervicovaginal shedding

The proportion of visits attended was 97.8%, 97.0%, and 94.8% in the three groups (table 2). The proportion of women who experienced at least one GUD episode over 24 weeks increased from 22.7% among HIV-uninfected women to 33.3% among women taking HAART and 42.6% among those not taking HAART (p = 0.22). Using a “per-visit” analysis, we observed a significantly higher proportion of visits with GUD among women not taking HAART (7.2%) compared with women taking HAART (3.1%) and HIV-uninfected women (1.9%) (p = 0.002).

Table 2 Clinical and virological outcomes during follow-up visits by study group

Of 72 GUD cases, 15 were vesicles and 18/37 (49%) swabbed ulcers had detectable HSV-2 DNA. The overall prevalence of herpetic ulcers was thus 63% (33/52).

The detection of cervicovaginal HSV-2 DNA at least once was more frequent among HIV-infected women taking HAART (63.3%) or not (67.6%), compared with HIV-uninfected women (45.5%) (p = 0.17; table 2). The proportion of women with detectable cervicovaginal HSV-2 DNA at each visit is shown in fig 1. The frequency of HSV-2 shedding was lowest among HIV-uninfected women and highest among untreated HIV-infected women. Using a “per-visit” analysis (table 2), we observed a significantly higher proportion of visits with detectable HSV-2 shedding among women not taking HAART (15.5%) compared with women taking HAART (9.7%) and HIV-uninfected women (4.3%) (p<0.001). When HSV-2 shedding was detected, the mean quantity of genital HSV-2 DNA was high but was similar across groups (∼4.6 log10 copies/ml).

Figure 1 Proportion of women with genital HSV-2 DNA at each visit, according to HIV-1 and HAART status. HAART, highly active antiretroviral therapy; HSV, herpes simplex virus..

Determinants of HSV-2 cervicovaginal shedding among HIV-infected women

The sociodemographic, clinical, or biological determinants of HSV-2 cervicovaginal shedding were assessed among HIV-infected women only (groups 2 and 3), because few HIV-uninfected women had any HSV-2 shedding episodes.

In both univariable and multivariable analyses (table 3), older age (adjusted risk ratio (aRR) 0.63, 95% CI 0.40 to 0.99) and vaginal washing (aRR 0.60, 95% CI 0.40 to 0.91) were associated with reduced HSV-2 genital shedding, whereas the presence of GUD was associated with an increased risk of HSV-2 cervicovaginal shedding (aRR 2.79, 95% CI 2.01 to 3.86). Compared with women not taking HAART and with CD4 cell counts of 500 cells/μl or greater, women on HAART had a similar risk of shedding (aRR 0.95, 95% CI 0.52 to 1.73), whereas women with CD4 cell counts of 200–500 cells/μl were more likely to shed HSV-2 (aRR 1.71, 95% CI 1.02 to 2.86). Bacterial vaginosis, other genital infections and STI syndromes were not associated with HSV-2 shedding.

Table 3 Determinants of HSV-2 cervicovaginal shedding among HIV-1 and HSV-2 co-infected women in Burkina Faso

Key messages

  • Compared with HIV-uninfected women, clinical and subclinical HSV-2 recurrences occur more frequently among HIV-1-infected women in Africa, particularly those with low CD4 cell counts, as previously shown in US studies.

  • Antiretroviral therapy appears only partly to reduce the frequency of genital herpes manifestations, but not to the levels of HIV-uninfected women.

  • HSV therapy may benefit HIV-infected individuals during HAART by reducing clinical recurrences and the transmissibility of both HSV-2 and HIV-1.

DISCUSSION

To our knowledge, no previous prospective study has described the natural history of genital herpes in sub-Saharan Africa including both clinical and virological outcomes, by HIV serostatus and antiretroviral therapy use. Most data on the natural history of genital herpes have originated from the United States. In one 2-month study conducted among 27 HIV-uninfected HSV-2-seropositive women with daily self-collection of genital swabs,25 genital HSV-2 shedding was detected by PCR on 28% of days, which contrasts with the 4.3% found in our study population. A recent study among 49 HIV-infected patients (only five women)14 has reported that genital HSV-2 DNA was detected on 29% of days, contrasting with 15.5% in our study. The same study also included 28 individuals (four women) treated with HAART in whom the proportion of HSV-2 shedding days was 14.9% contrasting with 9.7% in our study. The frequency of HSV-2 genital shedding among US women appears higher than among women in Burkina Faso, despite similar laboratory techniques with identical detection thresholds (500 copies/ml).6 14 21 22 Genital samples were, however, collected from several anatomical sites in US studies, including vaginal, vulvar and perianal areas, which increased the probability of detecting HSV-2 DNA, whereas our study relied on enriched cervicovaginal lavages only. Alternatively, differences may be attributable to the variable durations of HSV-2 infection. The frequency of HSV-2 shedding decreases quickly in the few years after HSV-2 infection.26 Differences in the age of HSV-2 acquisition and the age structure of study populations would thus probably influence the observed genital shedding rates. Our population of high-risk women may have acquired HSV-2 a long time ago, resulting in lower rates of HSV shedding, which is supported by the observed reduced shedding in women older than 34 years.

The occurrence of genital ulcerations followed the same trend as HSV-2 genital shedding, with a higher rate observed in HIV-infected women not taking HAART. The proportion of individuals experiencing clinical reactivations may be even higher in the general population of HIV-infected women as the study participants were not eligible for HSV-suppressive therapy.

A limitation of our study are the eligibility criteria used to enrol women in the randomised trials, which selected women with less frequently recurring herpes, hence perhaps the lower rates of shedding and genital ulcerations observed. Another limitation is its small sample size, and thus low power to detect differences in per-woman analyses. However, we had good power to detect differences between groups in the per-visit analyses. Despite these limitations in terms of generalisability to other African women, our findings underscore the need to improve HSV counselling, symptom recognition and management in Africa. Half the GUD cases were caused by HSV-2, in accordance with other African studies1820 and the Women’s Interagency HIV Study (a large US female cohort) study, which reported an almost doubling of the risk of GUD incidence among HIV-positive women compared with HIV-uninfected women.15

We found few correlates of HSV-2 genital shedding among HIV-infected women. Women with CD4 cell counts of 200–500 cells/μl were more likely to shed HSV-2 than women with higher CD4 cell counts, confirming previous findings13 27 and the theory that cell-mediated immunity is essential to control HSV-2 recurrences.28 We also found that women on HAART had a similar HSV-2 shedding frequency to women not on HAART with CD4 cell counts of 500 cells/μl or greater, but significantly higher than among HIV-uninfected women. This may be because women on HAART had not yet recovered their CD4 cell count, having been on antiretroviral drugs for a short period. However, Posavad et al14 suggested that a longer duration of HAART (median 21 months) could only decrease clinical but not subclinical HSV-2 reactivations. Similarly, the use of HAART and subsequent immune recovery did not completely eliminate the occurrence of clinical HSV reactivations in the Women’s Interagency HIV Study cohort.15 We suggest that the residual HIV genital replication observed in patients taking HAART with suppressed plasma levels21 29 might enhance HSV reactivations and vice versa, through direct or indirect viral interactions.

The inverse association between vaginal washing and HSV-2 shedding was unexpected and may be caused by the mechanical removal of vaginal secretions containing HSV-2 DNA. Although women were asked not to perform vaginal washing on their visit days, the washing effect from previous days may not be negligible. A longitudinal study among US women did not find any association between douching and HSV-2 shedding, although there was an association between shedding and bacterial vaginosis,24 and bacterial vaginosis has been associated with douching.30 Interestingly, we observed a trend for a role of bacterial vaginosis on HSV-2-DNA shedding. The lack of statistical significance may be due to the relatively low number of HSV-2 shedding episodes. Bacterial vaginosis has been found to enhance HSV-2 acquisition,31 and to be a significant correlate of genital HSV-2 shedding among immunocompetent women in the United States.24 Reciprocally, HSV-2 seropositive women may experience more frequent bacterial vaginosis episodes.32 The interactions between bacterial vaginosis and HSV-2 are important to decipher, because both infections may individually or synergistically impact on HIV transmission.

In conclusion, we found that clinical and subclinical HSV-2 reactivations occur more frequently among HIV-infected women, particularly those with lower CD4 cell counts. The use of HAART only partly reduces the frequency of HSV-2 genital shedding. HSV therapy may benefit HIV-infected individuals, including patients responding well to HAART.

Acknowledgments

The authors wish to thank the women and the organisations of persons living with HIV/AIDS who participated in this study (‘Yerelon’, ‘Espoir et Vie’, Centre ‘Solidarité Action Sociale’ and ‘Espoir pour Demain’); and staff at Service d’Hygiène, Bobo-Dioulasso. They would also like to thank David Mabey (London School of Hygiene and Tropical Medicine) for comments on an earlier draft of this manuscript.

REFERENCES

View Abstract

Footnotes

  • Funding: This study was sponsored by France’s Agence Nationale de Recherches sur le SIDA et les Hépatites (ANRS), grants ANRS-1291 and ANRS-1285. Additional financial support was provided through the UK’s Department for International Development (DFID)-funded Knowledge Programme on HIV/AIDS and STI of the London School of Hygiene and Tropical Medicine.

  • Competing interests: None declared.

  • Ethics approval: The research protocol was approved by the institutional review board at Centre Muraz, and by the research ethics committees at the Burkina Faso Ministry of Health and at the London School of Hygiene and Tropical Medicine. The trials were registered with the International Clinical Trials Registers (NCT00158509). Eligible women provided witnessed written (or thumb-printed) consent. Study procedures complied with the principles of the Helsinki Declaration and the ANRS Research Ethics Charter.

  • Contributors: PM, NN, HAW and PVP conceived the study, which was carried out by NN, IK, AO and AS. Laboratory analyses were performed by MCD in Bobo Dioulasso and VF and MS in Montpellier. Statistical analyses were performed by NN and HAW and reviewed by PM. The first manuscript was drafted by PM and NN with inputs and reviews from all co-authors who approved the final version of the manuscript.

  • Composition of the ANRS 1285 Study Group: Eloi Bahembera, Abdramane Berthé, Minata Coulibaly, Marie-Christine Defer, Ramata Diallo, Didier Djagbaré, Issouf Konaté, Florent Ky-Dama, Gilles T M’Boutiki, Nicolas Méda, Inès Millogo, Nicolas Nagot, Abdoulaye Ouedraogo, Djénéba Ouedraogo, Francois Rouet, Anselme Sanon, Haoua Sawadogo, Roselyne Vallo and Laurence Vergne (deceased January 2007) (Centre Muraz, Bobo-Dioulasso, Burkina Faso); Philippe Mayaud, Nicolas Nagot and Helen A Weiss (London School of Hygiene and Tropical Medicine, London, UK); Pierre Becquart, Vincent Foulongne, Michel Segondy and Philippe Van de Perre, (Université Montpellier 1 and CHU Montpellier, Montpellier, France); Jean-Baptiste Andonaba and Adrien Sawadogo (University Hospital of Bobo-Dioulasso, Burkina Faso).

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Brief encounters
    Helen Ward Rob Miller