OBJECTIVES: Several studies have demonstrated an association between herpes simplex virus type 2 (HSV-2) and HIV-1, but available data on risk factors for HSV-2 acquisition are limited. Overall prevalences for HSV-1 and -2 were 56% and 19%, respectively. They provided interview information; blood for HSV-2, HIV and syphilis serology; first void urine specimens for diagnosis of Neisseria gonorrhoeae and Chlamydia trachomatis infection; and genital specimens for diagnosis of bacterial vaginosis, vaginal candidiasis and trichomoniasis. At baseline and every 3 months thereafter, participants were interviewed, and blood and genital samples were collected. Association of HSV-2 seropositivity with socio-demographic and behavioral characteristics and HIV status was analyzed. Patients already on antiretroviral therapy (ART) were excluded. A higher HSV-2 seroprevalence was associated with female sex, older age, marital status (with single subjects been under the greater risk), irregular condom use on sexual intercourses with new partners and with higher number (five or more) of sexual partners during lifetime.
Factors associated with HSV-2 infection were methamphetamine use (OR = 2.30, 95%CI:1.43-3.70), having first commercial sex intercourse at under 20 years of age (OR = 1.77, 95% CI:1.16-2.69), working at low-end establishments (OR = 2.61, 95%CI:1.66-4.10), working at the local sex-work venues for more than one year (OR = 2.01, 95% CI:1.28-3.14), with low education backgrounds (OR = 2.27, 95% CI:1.18-4.36), using methamphetamine (OR = 3.95, 95% CI:2.09-7.44), low education background (OR = 2.43, 95% CI: 1.10-5.36), 21-30 year-olds (OR = 2.95, 95%CI:1.08-8.03), older than 31 years of age (OR = 7.05, 95% CI:2.48-20.01) etc., were independent risk factors associated with the superinfection of syphilis and HSV-2. Donner, and J. HSV-2 and HIV seropositivity are strong markers for high-risk sexual behavior. The 23% annual HSV-2 incidence in this study is among the highest reported anywhere in the world. New immunologic assays that use type-specific glycoproteins such as gG1 and gG2 distinguish herpes simplex type 2 (HSV-2) from type 1 (HSV-1) infection and permit better understanding of the epidemiology of genital herpes [1, 2]. Demographic characteristics associated with increased HSV-2 prevalence in the United States include female sex, race, older age, lower education level, and lower income [1–4]. HSV-2 seropositivity is also correlated with markers of sexual exposure, such as lifetime number of partners, years of sexual activity, sexual orientation, and history of other sexually transmitted disease (STD) [3, 5, 6].
A limited number of longitudinal studies in the United States find HSV-2 incidence ranging from 2% over 4 years for college students to 5% over 6 months among STD clinic patients . The prevalence and incidence of HSV-2 in sub-Saharan African populations are of particular concern in light of the potential causal association between genital herpes and transmission of human immunodeficiency virus (HIV) [7–11]. Studies on the African continent show a wide variation in HSV-2 prevalence across diverse populations: 20% among surgical patients in Senegal, 28% among Rwandan military recruits, 36% among patients with genital ulcer disease (GUD) in Uganda, 41% among adults in Congo (Kinshasa), 43% among patients with STDs in Tanzania, 60% among men older than 30 years in rural Tanzania, 68% in rural Uganda, 71% among adults in Congo (Brazzaville), and 96% among prostitutes in Senegal [8, 12–15]. Measures of HSV-2 incidence in sub-Saharan Africa are rare; 1 study in rural Uganda found an incidence of 16.7% over 12 months . The greatest risk was associated with renting. We report here the prevalence of HSV-2 infection upon enrollment in the study, the incidence of HSV-2 infection during follow-up, and risk factors for the presence and acquisition of HSV-2 infection. We further assess the temporal relation between HSV-2 and HIV seroconversion.
Stored sera screened for HSV-2—specific antibodies originated from male factory workers enrolled in the Zimbabwe AIDS Prevention Project (ZAPP), a longitudinal study established to determine the prevalence, incidence, and correlates of HIV infection and to evaluate a peer education HIV prevention intervention [16–18]. Subjects were recruited and followed up at 40 factories in greater Harare, Zimbabwe, from March 1993 to June 1997. At enrollment and ∼6-month intervals, workers were interviewed on HIV-risk-related behaviors and blood was drawn for HIV testing. Risk factors for HIV acquisition have been described elsewhere . In June 1994, factories were randomized to HIV counseling and testing (control) or HIV counseling and testing plus a peer education program (intervention). Subjects in the present analysis were male factory workers who had baseline and follow-up serologic testing prior to 14 June 1997 and who had sufficient stored sera for HSV-2-specific antibody testing. Of the 3381 subjects ever enrolled in the cohort, 2397 (71%) met these criteria.
Age, marital status, education, history of genital ulcer, history of STD, and number of sex partners in the last year did not differ between subjects included in the present analysis and those not included. However, subjects screened for HSV-2 had lower monthly income compared with those not screened (Z$734 vs. Z$948, t test P < .001), were less likely to report prostitute contact (18.6% vs. 23.8%, χ2 test P = .001), and had a lower prevalence of HIV at enrollment (16.6% vs. 27.5%, χ2 test P < .001). HIV test results were available to subjects through trained counselors within 2 weeks of each blood draw. However, HSV-2 testing was done after study end and results were not available to subjects. A strip recombinant immunoblot assay (RIBA HSV type 1/type 2 SIA; Chiron, Emeryville, CA) was used to detect and differentiate HSV-1— and HSV-2—specific antibodies [9, 19]. The RIBA type 1/type 2 nitrocellulose strips include recombinant antigen bands from HSV-1 (gG1 and gB1) and HSV-2 (gG2 and gD2) as well as controls for IgG. Antibodies specific for HSV-1 will react with gG1 and gB1 antigen bands but not with the gG2 band. HSV-2 antibodies will react with gG2 and gD2 bands but not with gG1 and gB1 bands. Compared with Western blot, the sensitivity of the RIBA assay is 95.1% for HSV-1 and 98.2% for HSV-2; specificity compared with Western blot is 99.4% for both HSV-1 and HSV-2 . The presence of HIV antibodies was demonstrated by the use of a third-generation enzyme immunoassay (EIA; HIV-1/HIV-2; Abbott Laboratories, Abbott Park, IL). Specimens reactive or indeterminate in the Abbott EIA were retested with a second third-generation EIA (Enzygnost Anti-HIV 1/2 Plus; Behring, Marburg, Germany). Samples were considered HIV antibody positive when positive results were obtained from both EIAs. Indeterminate or conflicting results were resolved by Western blot (HIV Blot 2.2; Diagnostic Biotechnology, Singapore). HSV-2 prevalence was estimated as the number of subjects testing HSV-2 seropositive at enrollment divided by the number of enrollment specimens screened. HSV-2 incidence was calculated among subjects testing HSV-2 seronegative at enrollment by dividing the number of seroconversions by the person-years of observation. For subjects remaining HSV-2 seronegative, person-years (py) of observation were calculated as the interval of time from enrollment to the most recent date of follow-up. For subjects seroconverting for HSV-2, person-years were calculated as the interval of time from enrollment to the midpoint between the last HSV-2 seronegative test and the first HSV-2 seropositive test. Ninety-five percent confidence intervals (CIs) for incidence estimates were based on a Poisson distribution. HSV-2 prevalence and incidence were calculated for subpopulations defined by demographic characteristics and reported risk factors relating to sexual exposure. Correlates of HSV-2 prevalence were identified by use of logistic regression analysis. Correlates of HSV-2 and HIV incidence were assessed by Cox proportional hazards analysis. To determine independent associations with HSV-2 acquisition, we first considered variables separately in a series of bivariate analyses. All variables significant in bivariate analysis at P < .05 were entered into a full model; those remaining significant were retained in the final model. The continuous variables of age, income, and number of sex partners were assessed as ordered categories. Odds ratios and hazard ratios were calculated for each stratum separately and for the linear trend across increasing strata. The final multivariate models assess the linear trends across increasing strata. The temporal sequence of HIV and HSV-2 acquisition was assessed by comparison of the dates of seroconversions for subjects who had serologic evidence of both infections. The number of cases in which HSV-2 seroconversion preceded HIV seroconversion was compared with the number of cases in which HIV seroconversion preceded HSV-2 seroconversion by use of the Wilcoxon matched-pairs sign-rank test. Other statistics were used as indicated in the text. Of the 2397 male Zimbabwean factory workers screened, 953 (39.8%) were HSV-2—positive at enrollment in the longitudinal study. History of genital ulcer in the year preceding enrollment was reported by 7.0% of HSV-2 seropositive workers and by 2.4% of HSV-2 seronegative workers. Table 1 presents HSV-2 prevalence by demographic characteristics and reported risk factors. HSV-2 prevalence increased with increasing age, beginning at 6.5% among workers aged 18–20 years and reaching 62.2% among workers 46 years and older. Of all groups examined, HSV-2 prevalence was highest (76.4%) among workers who were also HIV positive at enrollment. Paradoxically, workers with lower education but higher income had higher prevalence of HSV-2; however, the association of lower education and HSV-2 seropositivity disappeared after we controlled for age and income. In multiple logistic regression analysis (table 2), HSV-2 seropositivity was significantly associated with HIV seropositivity (adjusted odds ratio [OR], 7.2; 95% CI, 5.4–9.4), marital status (divorced OR, 4.1; 95% CI, 2.0–8.6; married, reside apart OR, 3.0; 95% CI, 2.2–4.1; married, reside together OR, 2.9; 95% CI, 2.3–3.9), history of any STD in the previous year (OR, 2.1; 95% CI, 1.5–2.9), older age (OR, 1.4 per age group; 95% CI, 1.3–1.5), and higher income (OR per quartile, 1.1; 95% CI, 1.0–1.2). The incidence of HSV-2 was 6.2/100 py (95% CI, 5.4–7.0), based on 204 seroconversions among the 1444 men who enrolled seronegative and observed for 3316 py of follow-up. History of genital ulcer during the follow-up period was reported by 15.2% of workers who seroconverted for HSV-2 and by 3.2% of workers who did not seroconvert. Of the 51 workers reporting genital ulcers during the follow-up period, 31 (57.4%) seroconverted for HSV-2. Table 3 displays HSV-2 incidence by demographic characteristics and reported risk factors. Unlike HSV-2 prevalence, HSV-2 incidence did not differ significantly by age, education, or income. Also in contrast to HSV-2 prevalence, single men did not have a lower incidence of HSV-2 compared with married men. Widowers had the highest incidence of HSV-2 (50.0/100 py); the estimate, however, is based on 2 observed seroconversions (95% CI, 6.1–180.5/100 py). HSV-2 incidence was 23.0/100 py (95% CI, 16.7–31.0) among workers who enrolled HIV-positive and 24.8/100 py (17.4–34.4) among workers who seroconverted for HIV during the follow-up period. Variables independently associated with HSV-2 acquisition in multiple Cox proportional hazard analysis (table 4) were being widowed (adjusted hazard ratio [HR], 5.4; 95% CI, 1.3 ± 22.0), HIV seropositive at baseline (HR, 4.7; 95% CI, 3.3 ± 6.7) or HIV seroconverting during follow-up (HR, 3.9; 95% CI, 2.6 ± 5.8), history of genital ulcer (HR, 2.0; 95% CI, 1.2 ± 3.3), history of any STD (HR, 1.9; 95% CI, 1.3 ± 2.9), and number of sex partners (HR per partner, 1.1; 95% CI, 1.0 ± 1.3). HSV-2 incidence was not independently associated with age, education, income, paying for sex, or working at a factory with a peer education HIV prevention program.
Incidence of herpes simplex type 2 (HSV-2) among male factory workers during follow-up in a cohort, by demographic characteristics and risk factors, Harare, Zimbabwe, March 1993–June 1997 (n = 1444). Figure 1 displays acquisition of HSV-2 and HIV for 4 groups defined by serostatus at enrollment. Among workers negative for both infections at enrollment, 125 seroconverted for HSV-2 alone, 36 seroconverted for HIV alone, and 36 seroconverted for HSV-2 and HIV during follow-up. Among workers positive for HSV-2 but negative for HIV at enrollment, 52 seroconverted for HIV. Among workers negative for HSV-2 but positive for HIV at enrollment, 43 seroconverted for HSV-2. The incidence of HIV among workers who were HSV-2-positive at enrollment was 3.0/100 py, compared with 2.1/100 py among those who were HSV-2—negative at enrollment (P = .059, assuming Poisson distribution). The incidence of HSV-2 among workers HIV-positive at enrollment was 23.0/100 py compared with 5.1/100 py among those HIV-negative at enrollment (P < .001, assuming Poisson distribution). Preexisting infection with HSV-2 at enrollment (HR, 3.5; 95% CI, 2.2 ± 5.8) and HSV-2 seroconversion during follow-up (HR, 6.7; 95% CI, 4.2 ± 10.7) were strong predictors of HIV acquisition even after we controlled for age, marital status, education, income, history of genital ulcer, history of STD, number of sex partners, and prostitute contact. The temporal relation of HSV-2 and HIV acquisition was examined for the 435 workers with serologic evidence of both infections by study end. For 317 workers, the temporal sequence of seroconversions could not be determined because they either enrolled seropositive for both infections (n = 304) or were found on the same follow-up visit to have seroconverted for both infections (n = 13). Of the remaining 118 dually infected workers, 61 acquired HSV-2 first (52 who enrolled HSV-2—positive plus 9 seroconversions) and 57 acquired HIV first (43 who enrolled HIV-positive plus 14 seroconversions). The data did not provide evidence that 1 infection was more likely to precede the other whether we examined all workers with both infections (P = .830, sign-rank test) or only those who were observed to seroconvert for both infections during follow-up (P = .880, sign-rank test). History of genital ulcer was examined by baseline and follow-up HIV serostatus. Of 304 workers seropositive for both infections at baseline, 41 (13.5%) reported history of genital ulcer in the preceding year, compared with 26 (4.0%) of 649 workers who were HSV-2—positive but HIV-negative (P < .001, χ2 test). Of 36 workers who seroconverted for both HSV-2 and HIV, 13 (36.1%) reported a genital ulcer during follow-up, compared with 18 (14.4%) of 125 who seroconverted for HSV-2 only (P = .004, χ2). The prevalence of HSV-2 infection in this general population of male factory workers was high, 39.8%, but plausible given the wide range of estimates for sub-Saharan Africa. Not surprisingly, a strong association between HSV-2 and HIV seropositivity was found, consistent with cross-sectional seroepidemiologic surveys on the African continent [8, 13, 15]. HSV-2 seropositivity was also associated with ever being married, history of STD, older age, and higher income. The latter finding is in contrast to the United States, where lower income is associated with increased HSV-2 prevalence [2, 4]. A possible explanation is that money may afford increased access to sexualpartners, particularly prostitutes, to a greater degree among this urban working population in sub-Saharan Africa. The observed rate of HSV-2 seroconversion was also high, 6.2/100 py. The estimate is comparable to populations in the United States at high risk for STD, such as multipartnered heterosexual STD clinic patients in Atlanta (5% over 6 months) and gay men in San Francisco (5%/year) . One longitudinal study from the African continent (rural Uganda) estimated HSV-2 incidence among men at 11.8% over 12 months; however, the sample size was relatively small (2 seroconversions among 17 men, 95% CI, 1.4%–42.5%) . Thirty-one (15.2%) of 204 workers seroconverting for HSV-2 gave a history of genital ulcer during follow-up and 31 (57.4%) of 54 workers reporting genital ulcers were observed to seroconvert. The latter finding suggests that HSV-2 is the most common cause of genital ulcers in this population. In contrast to HSV-2 prevalence, the incidence of HSV-2 was surprisingly constant across age, marital status (except widowers), education, and income groups. The difference is likely explained by HSV-2 seropositivity reflecting lifetime risk, whereas HSV-2 seroconversion reflects exposure during the study period. Enrollment while HIV-positive or acquisition of HIV during follow-up remained significantly associated with HSV-2 seroconversion after we controlled for other markers of sexual exposure. A discouraging finding was that the presence of a workplace peer education program, which included on-site availability of free condoms and promotion of their use, was not associated with reduced HSV-2 incidence. The hypothesis that GUD is causally associated with the acquisition of HIV infection has strong appeal. Genital ulceration may enhance acquisition of HIV by disrupting mucosal membranes, providing an easy portal of entry [20, 21]. Several studies, including the present one, have demonstrated a strong correlation between HIV and HSV-2 infection even after controlling for behaviors likely to lead to both infections [7–11]. Nevertheless, whether HSV-2 infection causes a predisposition to acquisition of HIV is debated [21–24]. Studies cannot completely control for behavior and are vulnerable to biases of self-reported information. In fact, HSV-2 serology is proposed as a more objective marker for sexual behavior than self-report [5, 12]. Moreover, precise measures of exposure, such as episodes of sexual intercourse with infected partners, are usually not available. Last, many studies were not longitudinal or could not establish whether HSV-2 infection preceded HIV acquisition, a key criterion for causality. At least 2 studies that could determine the temporal sequence of infections found no association between prior HSV-2 infection and HIV seroconversion [21, 24]. The present longitudinal study found that prior HSV-2 infection was associated with a 1.4-fold increase in the rate of HIV acquisition. In contrast, prior HIV infection was associated with a 4.5-fold increase in the rate of HSV-2 acquisition. Moreover, among workers whose seroconversions for both infections were observed during follow-up, more seroconverted for HIV before HSV-2 than vice versa. Overall, however, there was no statistical evidence in this cohort that 1 seroconversion was more likely to precede the other. The evidence would support that either infection predisposes individuals to the other, or, more simply, that persons with either infection are continuing to engage in unsafe sex. A third explanation is that infection causes increased transmission to others rather than increased susceptibility. Higher incidence of HSV-2 among persons with HIV infection may be observed because people who are HIV positive are more likely to continue to have HIV-positive partners. Because HIV-positive individuals with HSV-2 shed virus more often [25–27], such partners may be more infectious and pose a higher risk of transmission. In this scenario, seropositivity may be a marker for increased infectiousness of sexual partners as opposed to increased susceptibility to infection. Coinfection may enhance transmission of HIV and HSV-2 to others through increased reciprocal viral replication, increased viral shedding, and increased frequency of genital ulceration as a portal for viral excretion. In this cohort, we found a >3-fold increase in reported genital ulcers in the previous year among HSV-2—positive workers who enrolled while HIV positive and a >2-fold increase in genital ulcers among workers who seroconverted for both infections during follow-up.
Other studies have demonstrated increased replication of HIV in herpes virus-infected individuals [25, 26] and increased shedding of HSV-2 among HIV-seropositive women compared with seronegative women . Unfortunately, without data on sexual partners in the present study, we cannot draw definitive conclusions. We recognize other limitations to our data. We were unable to determine the timing of seroconversions for the vast majority of persons with both infections (317 of 435). Precise measures of sexual behavior, sexual exposures, period of active genital lesions, viral replication, viral shedding, and physical examinations were lacking. Moreover, differences between cohort subjects who were included in the study of HIV but not HSV-2 may affect inference on the relation between these infections. Notably, subjects screened for HSV-2 had lower incomes, less prostitute contact, and lower prevalence of HIV.
Whether HSV-2 and HIV transmission are linked through increased susceptibility or through increased infectiousness, the high incidence of both infections in this cohort is evidence of continued high-risk behavior likely to fuel both epidemics. Improved interventions and prevention programs targeted to populations in which the incidence of either infection is high are needed. We acknowledge the contribution of the investigators and staff of the Zimbabwe AIDS Prevention Project (in alphabetical order): B. Brown, P. Chikukwa, M. Jeche, A. Mashingaidze, C.
Maposhere, P. Mason, M. Matshaka, M. Mbizvo, L. Moses, D. Mvere, V. Mzezewa, S.
Ray, L. Rogers, H. Smith, O. Tobaiwa, and S. Whytehead. The study was carried out as part of HIVNET 011 for which written informed consent was obtained from all participants. Human experimentation guidelines of the US Department of Health and Human Services, the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, Stanford University Medical Center Institutional Review Board, and the Medical Research Council of Zimbabwe were followed.
Grant support: US Public Health Service (award A1 33868, Preparation for AIDS Vaccine Evaluation) and the HIV Prevention Network, Family Health International (subcontract A1-36173-129). HSV1+2 assays were donated by Chiron Corporation (Emeryville, CA).