Herpes simplex virus type II (HSV-2) infection is the most common cause of genital ulcer disease in the world1,2 and may increase the risk of HIV-1 transmission to women through disruption and inflammation of the epithelial barrier.3 Recent studies reveal that even in the absence of genital ulcer disease, there is a correlation between genital tract shedding of HIV-1 and HSV-2.4,5 In sub-Saharan Africa, where HIV-1 infection is spread mainly by heterosexual transmission, other sexually transmitted infections, including HSV-2 infection, are common.6-8 In South Africa in 1994, 50% of men attending sexually transmitted disease clinics were infected with HSV-2, which was the most common cause of genital ulcer disease,9 while a community survey carried out in 1999 in Carletonville found that 89% of women and 42% of men were infected with HSV-2 by the time they reached 25 years of age.10 Among sex workers and other high-risk groups, high rates of both HSV-2 and HIV-1 infections have been reported in Africa and elsewhere.11,12
Several studies have found a positive association between HSV-2 infection and HIV-1 infection.7,13-15 In a study conducted in 4 African cities, the adjusted odds ratio for prevalent HIV-1 and HSV-2 infection among women ranged from 4.0 in Kenya to 5.5 in Yaounde.12 In this paper we estimate the extent to which prevalent HSV-2 infections increase the risk of acquiring HIV-1 infections among a group of female commercial sex workers in KwaZulu-Natal, South Africa, and whether this risk is increased further for women with recently acquired HSV-2 infections. In 3 studies1,16,17 estimates have been made of the hazard or odds ratio for HIV-1 in people who were HSV-2 positive compared with people who were HSV-2 negative throughout the study (we refer to this as the hazard or odds ratio for prevalent infection) and the hazard or odds ratio for HIV-1 in people who seroconverted to HSV-2 during the study, regardless of when they seroconverted, as compared with people who were HSV-2 negative throughout the study (we refer to this as the “hazard or odds ratio for incident infection”).
In the first study 2397 male factory workers in Zimbabwe16 were followed up every 6 months for 4 years. HSV-2 incidence was 6.2% per year while HIV-1 incidence was about 2.5% per year. The adjusted HIV-1 hazard ratios for prevalent and incident HSV-2 infection were 3.5 (2.2-5.8; unless otherwise stated errors are 95% CIs) and 6.7 (4.2-10.7), respectively. The adjusted HSV-2 hazard ratios for prevalent and incident HIV-1 infection were 4.7 (3.3-6.7) and 3.9 (2.6-5.8), respectively. There was no evidence that either infection was more or less likely to precede the other.
In the second, case-control study in rural Tanzania17 with 70 male and 57 female cases (and 636 controls), HSV-2 incidence was about 10% per year (but 20% per year in women <20 years old) while HIV-1 incidence was about 0.75% per year.18 Among men the adjusted HIV-1 odds ratios for prevalent and incident HSV-2 infection were 6.1 (2.5-14.9) and 16.8 (6.1-46.3), respectively. These rates were higher than, but not significantly different from, the hazard ratios found in Zimbabwe. For women the effect of HSV-2 infection on the incidence of HIV-1 was not significant and the corresponding adjusted odds ratios were 1.3 (0.6-2.8) and 2.4 (0.8-6.8), respectively.
In the 3rd study in Pune, India,1 2269 male patients with sexually transmitted infection, 9 of whom were “hijra” (eunuchs), and 463 women who were partners of male sexually transmitted infection (STI) patients, commercial sex workers, or had reproductive tract infections, were followed up 3 times over 11 months (median values). The HSV-2 incidence was 11.4% (9.9%-13.0%) per year while the HIV-1 incidence was 5.8% (5.0%-6.6%) per year. In this study, which included mainly men, the HIV-1 adjusted hazard ratios for prevalent and incident HSV-2 infection were 1.7 (1.2-2.3) and 2.7 (1.7-4.3), respectively. (The latter estimate is the weighted average of “recent” and “remote” incident infections as given by the authors). In all 3 studies, the HIV-1 hazard or odds ratios for HSV-2 incident infection were greater than for HSV-2 prevalent infection. However, in none of these studies were the differences in the hazard or odds ratios significant. Furthermore, in the one study for which results were reported separately for women,17 neither odds ratio differed significantly from 1.
Here we investigate the impact of prevalent and incident HSV-2 infection on the incidence of HIV-1 infection in a cohort of female sex workers who were followed up and tested monthly for antibodies to HIV-1 and HSV-2. The HSV-2 incidence was 35% per year while the HIV-1 incidence was 18% per year. The longest follow-up was for 3.6 years (median 2.2 years). In addition we do an analysis with HSV-2 seroconversion treated as a time-dependent covariate. We provide a comparative analysis of the previously published studies to determine more precisely the hazard ratio for HIV-1 incidence among people with prevalent and incident HSV-2 infections.
STUDY POPULATION AND METHODS
A total of 416 sex workers from 5 truck stops between the port city of Durban and the commercial center of Johannesburg, South Africa, were invited to participate in a multicenter vaginal microbicide (Advantage S: Columbia Laboratories, Paris, France, containing 52.5 mg nonoxynol-9, N9) phase 3 clinical trial funded by UNAIDS, which has been described in detail elsewhere.19
Of the 416 women screened, 208 were excluded because they were already HIV-1 positive. Because the original study was a trial of a microbicide that was still undergoing reproductive toxicology trials, and in which condoms were promoted, a further 12 women were excluded because they had <5 sexual partners per week, were pregnant or planning to become pregnant, or were allergic to latex. At enrollment the 196 eligible HIV-negative women were educated in HIV-1 prevention, counseled on safe sexual behavior, and randomly assigned to either the N9 or the placebo arm.20 Nine women did not return after the enrollment visit and were not included in the survival analysis. During each follow-up visit the women were counseled and told that other STIs may increase their likelihood of being infected with HIV, and blood was drawn to test for HIV-1 infections. The HSV-2 tests were done retrospectively after the main study had been completed. A clinician gave each woman a gynecologic examination and took swabs to test for Trichomonas vaginalis, Chlamydia trachomatis, Neisseria gonorrhoeae, and candidiasis. STIs were treated according to the South African syndromic management guidelines.
At each follow-up visit, the women completed a questionnaire on sexual behavior, were provided with condoms, and were counseled to use them whenever they had sex. The study was approved by the University of Natal Ethics Committee and the agencies that funded the initial trial.
Blood specimens were tested for HIV-1 infection using a highly sensitive antibody enzyme-linked immunosorbent assay (ELISA) (Vironostika) and positive specimens were retested using a 2nd highly specific ELISA (Abbott). Recently acquired syphilis was determined using an rapid plasma reagin (RPR) test confirmed with treponema pallidium haemagglutination assay/fluorescent treponemal antibody test (TPHA/FTA) tests. T. vaginalis and candidiasis were tested by wet mount. Endocervical swabs were taken and tested for C. trachomatis using an ELISA (Syva). N. gonorrhoeae infection was detected by culture. HSV-2 was detected using Gull ELISA (Meridian) for which the sensitivity was 93% and the specificity was 97%.21,22 All HIV-1 seroconversions were confirmed at the Institute of Tropical Medicine, Antwerp, Belgium. Once the HIV-1 infection was confirmed, sera were tested for HSV-2 infection starting from the date of HIV-1 seroconversion and going back to the first positive HSV-2 test.
Analyses were done using Stata (Stata Corporation, College Station, TX) (Release 8). Kaplan-Meier survival curves for the time to HIV-1 seroconversion were calculated for various groups of women and we estimated crude incidence rates for each group. Three Cox proportional hazards models were used to determine hazard ratios for HIV-1 seroconversion and to allow for confounding variables. Model I was used to test potential confounding variables for inclusion in subsequent models. Model II was used to test for the effect of HSV-2 admission status on entry and exit while adjusting for confounders. Model III was used to test for the effect of recent HSV-2 seroconversion by including it as a time-dependent variable in the Cox proportional hazards model. Schoenfeld residuals were used to test the assumption of proportional hazards.
The prevalence of both HIV-1 and HSV-2 among these women was high. On recruitment, 50% of 416 women tested positive for HIV-1 infection and 84% for HSV-2 infection, and HIV-1 and HSV-2 serostatus were strongly associated (odds ratio = 4.6, 2.5-8.3). As in other populations in South Africa, the prevalence of HIV-1 increases with age, peaks among women at the age of about 24 years, and declines with age thereafter.23 The prevalence of HSV-2 infection, conversely, increases rapidly with age, reaching 60% at the age of 18 years, 90% at the age of 25 years, and 100% by the age of 35 years (Fig. 1).
Baseline data for potential risk factors and the prevalence of HSV-2 and HIV-1 are given in Tables 1 and 2. Education levels were reasonably high; 49% of the women had 4-7 years of education while a further 39% had ≥8 years of education. Reported condom use was low and 41% of women said that they never used condoms, while only 11% said that they always used condoms. Forty percent of women had engaged in anal sex. Curable sexually transmitted infections were common, with 31% testing positive for syphilis, 10% for N. gonorrhoeae, 12% for C. trachomatis, and 36% for T. vaginalis infection. The average age of the women was 25 years; they had an average of 20 partners per week and had been doing sex work for an average of 2.5 years. Of the categorical variables (Table 1), syphilis was strongly associated with HSV-2 infection (P = 0.004) while C. trachomatis infection was weakly associated with HSV-2 infection (P = 0.059). Of the continuous variables (Table 2), age was significantly associated with HSV-2 status (P = 0.001).
Figure 2A shows Kaplan-Meier curves for time to HIV seroconversion for women who were HSV-2 positive throughout (PP), HSV-2 negative throughout (NN), and HSV-2 negative on entry and positive on exit (NP). During the study period 24 women seroconverted to HSV-2 and, of these, 20 seroconverted to HIV-1 after HSV-2; 3 seroconverted in the same time interval; and 1 seroconverted to HIV-1 before HSV-2. Table 3 gives the crude HIV-1 incidences for each group (PP, NN, NP) and these were not significantly different. When time was measured from HSV-2 seroconversion, the incidence of HIV-1 was significantly greater than in those who were HSV-2 positive on entry.
Table 4 shows the results of a univariate analysis for potential risk factors for HSV-2 infection; only age and syphilis were significantly associated with HSV-2 status at entry. Cox proportional hazards models were then used to determine HIV-1 hazard ratios for the different groups of women according to their HSV-2 status. In model I (Table 5), the HIV-1 hazard ratio was assessed in relation to HSV-2 status at entry on its own and then in combination with each potential confounding variable in a series of bivariate analyses; only “study group” is close to being significant (P = 0.078). The effect of HSV-2 status on HIV-1 incidence remained significant when combined with each of the other variables except in the case of age, but then age was itself not significant. The hazard ratio for HSV-2 status on entry was little changed by the inclusion of each of the other variables.
In model II (Table 6) the women were categorized into 3 groups: those who were HSV-2 seropositive on admission, HSV-2 seronegative throughout, and HSV-2 seronegative on admission but seropositive on exit. All variables with a P value <0.2 in model I (study group, duration of sex work, and gonorrhoea) were included in a multivariate model to assess their combined effect on HIV-1 incidence. Since both age and syphilis were significantly associated with the HSV-2 infection status of the women (Table 4), these were initially included, but neither age nor syphilis was then significant (P = 0.775 and 0.419, respectively) and they were omitted from the final model. It is worth noting that the women included in the final analysis covered a narrow age range (median 24 years; interquartile range 19-29 years). In this analysis women with prevalent infections (PP), ie, those who were positive throughout the study, were taken as the reference group both because it was the largest group and because we wished to see whether women with incident infections are at significantly greater risk than women with prevalent infections. Model II (Table 6) shows that of women who remained negative throughout (NN) or who seroconverted during the study (NP) had hazard ratios that were greater than for women who were positive on entry (PP); in neither case was the difference significant.
Finally, in model III we proceeded as for model II but we treated HSV-2 seroconversion as a time-dependent covariate. In this analysis we combine all people who are HSV-2 negative up to the time when they either seroconvert to HSV-2 or leave the study in a group that we call NT. We then consider all those who seroconvert during the study in a 2nd group and measure HIV-1 incidence from the time of HSV-2 seroconversion and we call this group PT. The statistical comparison group is, again, PP, those who were HSV-2 positive throughout. Kaplan-Meier curves for these 3 groups are given in Figure 2B. In this analysis (Table 6), the hazard ratio for HIV-1 seroconversion for women who were HSV-2 negative (NT) was greater than for women who were HSV-2 positive throughout (PP), but again the difference was not significant. However, the hazard ratio for HIV-1 seroconversion in women after they seroconverted to HSV-2 (PT) is now 6.0 (2.6-14.0) times that for women who remained positive throughout (PP).
The data suggest, therefore, that immediately after HSV-2 seroconversion women experience a substantial, and statistically significant, increase in the risk of acquiring HIV-1 (when HSV-2 incidence is included as a time-dependent variable; model III in Table 6). Conversely, the data do not provide conclusive evidence for differences between the HIV-1 hazard ratio for what we have defined above as prevalent and incident HSV-2 infection (ie, comparing the groups PP, NN, and NP in Table 6, Model II).
Comparisons With Other Studies
The HIV-1 hazard ratios for incident and prevalent HSV-2 infection have been measured in 2 other studies, carried out in Zimbabwe16 and India,1 and the corresponding odds ratios have been measured in a study in Tanzania.17 The results of these studies are shown in Figure 3, where the final model in this paper (model II, Table 6) has been repeated using a logistic regression to determine the odds ratios. In the studies on men16,17 or mainly on men,1 both prevalent and incident HSV-2 infection increased the risk of acquiring HIV-1 infection significantly. In the present study (recalculated so that the comparison is with women who remained seronegative throughout) and the study on women in Tanzania,17 neither prevalent nor incident HSV-2 infection increased the risk of acquiring HIV-1 infection significantly. The data from the India study1 are for a heterogeneous group of men and women, and if we put this study aside the other 4 results in Figure 3 suggest that HSV-2 has less relative impact on the risk of HIV-1 infection in women than it does in men.
In none of these studies is the HIV-1 hazard ratio or odds ratio significantly different for prevalent and incident HSV-2 infections. However, on a nonparametric binomial test, the probability that 5 results all show that the hazard or odds ratios for incident infections is greater than for prevalent infections is 0.031. Furthermore, the Zimbabwe, India, and South African studies give a weighted average for the HIV-1 hazard ratio for incident as compared with prevalent HSV-2 infection of 1.85 (1.16-2.95; P = 0.010) while the Tanzania study and the South African study (using a logistic regression) give a weighted average for the HIV-1 odds ratio for incident as compared with prevalent HSV-2 infection of 2.37 (1.09-5.17; P = 0.031).
In a study in San Francisco, incident HIV-1 and HSV-2 infections among men who have sex with men (MSM) were correlated while prevalent infection with either virus was not significantly associated with subsequent infection with the other.24 However, the median time between clinic visits was 9 months and there were only 11 HSV-2 seroconversions among the MSM, all of whom seroconverted to HIV-1 in the same period. While the results of this study are consistent with the results reported, here the temporal sequence of events could not be established and the statistical power was low.
Many studies have shown a strong positive association between infection with HSV-2 and HIV-1; a direct causal link is more difficult to establish. Nevertheless, if the presence of either infection increases the incidence of the other it is likely that in populations in which HSV-2 is common HIV-1 will also spread rapidly, and it has been suggested that this might explain some of the regional variation in HIV-1 infection rates in sub-Saharan Africa.17 Some studies have found that HSV-2 infection increases HIV-1 incidence in men1,16,17 but not in women17 and in one study the HIV-1 hazard ratio for incident or prevalent HSV-2 infection is between about 4-5.16 All of these studies suggest that the impact of HSV-2 on HIV-1 is greater for those with incident than for those with prevalent HSV-2 infection, but in none of these studies is the difference statistically significant. Here we present data to show that the incidence of HIV-1 infection is significantly greater in women with incident HSV-2 infection (when HSV-2 incidence is analyzed as a time-dependent variable) than it is in women with prevalent HSV-2 infection. Furthermore, we are able to show that if we combine the results from all the studies, the hazard or odds ratios for HIV-1 infection is about twice as great in people with incident as opposed to prevalent infections (where incident infections refer to people who seroconverted during the study) and that this difference is then statistically significant.
The studies discussed in this paper cover a wide range of settings including male factory workers in Zimbabwe, men and women in rural Tanzania, a mixture of men and women with STIs in India, and commercial sex workers in South Africa. The background rates of infection also vary substantially. The incidence of HSV-2 is lower in Zimbabwe (∼6% per year) but much higher in South Africa (∼35% per year) than it is in India or Tanzania (∼10% per year). The incidence of HSV-2 is 2-3 times greater than the incidence of HIV-1 in all studies except in Tanzania, where it is about 7 times greater. Nevertheless, a consistent picture emerges of the impact on men and women and when comparing incident and prevalent HSV-2 infections.
In this study women were followed up every month, while they were either followed up at approximately 3-monthly intervals in the Indian study,17 approximately 6-monthly intervals in the Zimbabwe study,16 and after 2 years in the Tanzanian case-control study.17 The study reported here is the only one in which the follow-up interval is short enough to analyze incident HSV-2 seroconversions as a time-dependent covariate.
This study has some limitations. The small sample size limits the power of the study and the study was a retrospective study, relying on data collected for another purpose. Nevertheless, the results are statistically significant and consistent with previous studies. While age may be a confounder, the age range of the women was narrow (Table 2) and even though women who were HSV-2 positive on entry were significantly older than women who were HSV-2 negative on entry, age was not significant in combination with HSV-2 status on entry (Table 5). The experience of anal sex is common among these women and likely to be underreported, and this may explain the lack of significance of anal sex in this study. The presence or otherwise of herpetic ulcers, as observed by clinicians or self-reported, during each time interval could also have been recorded and may have helped to explain the apparent importance of recent HSV-2 infection as a risk factor for HIV-1 infection. The results could also be confounded by periods of high sexual activity during which the chance of being infected with HIV-1 and HSV-2 would both increase. However, only 2 women acquired both infections in the same time interval, suggesting that simultaneous infection, at least, is unlikely. Future studies should attempt to measure levels of sexual activity more precisely to separate behavioral effects on HIV-1 incidence from the biologic effects of HSV-2.
The key findings, from this and the other 3 studies discussed here, are as follows. First, prevalent or incident HSV-2 infection increases the incidence of HIV infection in men, but the range of estimates is large. The hazard ratio is 1.7 for prevalent infections in India1; the odds ratio is 17 for incident infections in Tanzania17 (Fig. 3) and rather uncertain. Second, prevalent or incident HSV-2 infection does not increase the incidence of HIV infection in women significantly (when women who seroconverted during this and the Tanzanian study17 are compared with those who remained HSV-2 negative throughout) but does increase the incidence of HIV infection in women significantly by 6.0 (2.6-14.0) times when incident infection is analyzed as a time-dependent covariate in this study.
These studies, taken together, suggest that immediately after HSV-2 seroconversion, the risk of acquiring HIV-1 increases by up to 6 times in women and probably by rather more in men. As time goes by the effect appears to wane and may eventually disappear altogether in women, although it appears to remain significant in men. It is also likely that HIV-1 infection increases the risk of acquiring HSV-2 infection.16 Here we have only considered the impact on HIV-1 infection of HSV-2 in the person being infected and not of HSV-2 infection in the person who is the source of the infection. Since people with HSV-2 secrete higher concentrations of HIV-13 and vice versa,25 this may further enhance transmission of both viruses.
While the high prevalence of HSV-2 infection in many developing countries may be an important determinant of HIV infection, clinical manifestations of HSV-2 infection are diverse, individuals are often asymptomatic and may not seek medical care,12 and many clinics do not test for HSV-2 infection routinely because they lack the facilities or fail to recognize the importance of the infection. Latent HSV-2 infections may reactivate and cause symptoms intermittently, rendering the carrier very contagious.26
It seems likely that HSV-2 is an important factor in determining the overall risk of HIV-1 infection, and the data reported here should help to guide the development of dynamical transmission models of HIV-1 epidemics that may help to reveal the extent and the importance of the synergy with HSV-2. Better data are still needed, although conducting such studies may now be more difficult with the advent of the widespread use of antiretroviral therapies. It is important that public health workers assess the extent to which symptomatic treatment of HSV-2 and, if it were available, a vaccine for HSV-2 may change the course of the HIV-1 epidemic.
The authors thank Prof. Richard Hayes, Dr. Charles Lacey, Dr. Anna Wald, Dr. Eline Korenrompe, Prof. Carolyn Williamson, and Dr. Lut van Damme for their comments. We are especially grateful to the COL 1492 participants for their participation in the COL 1492 trial.
1. Reynolds SJ, Risbud AR, Shepherd ME, et al. Recent herpes simplex virus type 2 infection and the risk of human immunodeficiency virus type 1 acquisition in India. J Infect Dis
2. Schomogyi M, Wald A, Corey L. Herpes simplex virus-2 infection: an emerging disease? Infect Dis Clin North Am
3. Wasserheit JN. Epidemiological synergy: interrelationship between human immunodeficiency virus infection and other sexually transmitted diseases. Sex Transm Dis
4. McClelland RS, Wang CC, Overbaugh J, et al. Association between cervical shedding of Herpes simples virus and HIV. AIDS
5. Mbopi-Keou FX, Gresenguet G, Mayaud P, et al. Interactions between herpes simplex virus type 2 and human immunodeficiency virus type 1 infection in African women: opportunities for intervention. J Infect Dis
6. Johnson AM, Laga M. Heterosexual transmission of HIV-1. AIDS
7. Auvert B, Ballard R, Campbell C, et al. HIV infection among youth in a South African mining town is associated with herpes simplex virus-2 seropositivity and sexual behaviour. AIDS
8. Auvert B, Buve A, Ferry B, et al. Ecological and individual level analysis of risk factors for HIV infection in four urban populations in sub-Saharan Africa with different levels of HIV infection. AIDS
. 2001;15(Suppl 4):S15-S30.
9. Chen CY, Ballard RC, Beck-Sague M, et al. Human immunodeficiency virus infection and genital ulcer disease in South Africa. Sex Transm Dis
10. Williams BG, Gouws E. The epidemiology of human immunodeficiency virus in South Africa. Philos Trans R Soc Lond B Biol Sci
11. Kilmarx PH, Limpakarnjanarat K, Mastro TD, et al. HIV seroconversion in a prospective cohort of female sex workers in Northern Thailand: continued high incidence among brothel-based women. AIDS
12. Weiss HA, Buve A, Robinson NJ, et al. The epidemiology of HSV-2 infection and its association with HIV infection in four urban African populations. AIDS
. 2001;15(Suppl 4):S97-S108.
13. Hook EW, Cannon R, Nahmias AJ, et al. Herpes simplex virus infection as a risk factor for human immunodeficiency virus infection in heterosexuals. J Infect Dis
14. Gwanzura L, McFarland W, Alexander D, et al. Association between human immunodeficiency virus and herpes simplex virus type 2 seropositivity among male factory workers in Zimbabwe. J Infect Dis
15. Wald A, Link K. Risk of human immunodeficiency virus infection in herpes simplex virus type 2-seropositive persons: a meta-analysis. J Infect Dis
16. McFarland W, Gwanzura L, Bassett MT, et al. Prevalence and incidence of herpes simplex virus type 2 infection among male Zimbabwean factory workers. J Infect Dis
17. Rodriguez M, Obasi A, Mosha F, et al. Herpes simplex virus type 2 infection increases HIV incidence: a prospective study in rural Tanzania. AIDS
18. Gilson L, Mkanje R, Grosskurth H, et al. Cost-effectiveness of improved treatment services for sexually transmitted diseases in preventing HIV-1 infection in Mwanza Region, Tanzania. Lancet
19. Ramjee G, Abdool Karim SS, Sturm A. Sexually transmitted infections among sex workers in Kwazulu-Natal, South Africa. Sex Transm Dis
20. Van Damme L, Ramjee G, Alary M, et al. Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in female sex workers: a randomised controlled trial. Lancet
21. Wald A, Ashley-Morrow R. Serological testing for herpes simplex virus HSV-1 and HSV-2 infection. Clin Infect Dis
22. Dodd RY, Notari EP, Stramer SL. Current prevalence and incidence of infectious disease markers and estimated window-period risk in the American Red Cross blood donor population. Transfusion (Paris)
23. Williams BG, Gouws E, Colvin M, et al. Patterns of infection: using age prevalence data to understand the epidemic of HIV in South Africa. S Afr J Sci
24. Turner KR, McFarland W, Kellogg TA, et al. Incidence and prevalence of herpes simplex virus type 2 infection in persons seeking repeat HIV counseling and testing. Sex Transm Dis
25. Mbopi-Keou FX, Gresenguet G, Mayaud P, et al. Genital herpes simplex virus type 2 shedding is increased in HIV-infected women in Africa. AIDS
26. Wald A, Zeh J, Selke S, et al. Reactivation of genital herpes simplex virus type 2 infection is asymptomatic HSV-2 seropositive persons. N Engl J Med
Keywords:© 2005 Lippincott Williams & Wilkins, Inc.
HIV; herpes simplex virus; AIDS; incidence; South Africa; sex worker