Secondary Logo

Journal Logo


Herpes simplex virus type 2 infection increases HIV incidence: a prospective study in rural Tanzania

del Mar Pujades Rodríguez, Maríaa; Obasi, Angelaa,b; Mosha, Frankc; Todd, Jima,c; Brown, Davidd; Changalucha, Johnc; Mabey, Davida; Ross, Davida,c; Grosskurth, Heinera,b; Hayes, Richarda

Author Information
  • Free



Sub-Saharan Africa continues to be the region of the world most severely affected by the HIV epidemic. Almost 70% of all HIV-infected individuals live in this region, and in 2000 an estimated 3.8 million of 5.3 million new infections occurred here. HIV began to spread rapidly in sub-Saharan Africa in the late 1970s and early 1980s, and most transmission occurs through heterosexual intercourse [1–5].

Today, HIV infection is well established throughout the region, but there are substantial geographical variations in HIV prevalence. The highest rates of HIV prevalence and incidence have been recorded in parts of eastern and southern Africa, whereas lower rates are generally observed in western Africa. A number of explanations have been offered for these geographical variations. Male circumcision is more common in the west, and there is accumulating evidence that this may be protective against HIV infection in men [6–11]. In addition, genital ulcer disease (GUD) appears to occur more frequently in the east and south, and this observation together with evidence from epidemiological studies that HIV transmission is greatly enhanced in the presence of GUD, suggest that GUD may play an important role in the rapid spread of HIV in some parts of Africa [3,12].

Herpes simplex virus type 2 (HSV2) is one of the most common sexually transmitted infections worldwide [13–15], and is a major cause of GUD. High seroprevalences of HSV2 have been reported in both developed and developing countries, with particularly high rates in some parts of Africa [12,13,16–19]. Biological and epidemiological studies have provided evidence supporting the potential in vivo interaction between HSV2 and HIV. This interaction has been shown to be bi-directional, HSV2 infection enhancing the transmission of HIV [20–22], while HIV infection appears to exacerbate the clinical expression of HSV2 [20,23,24].

The high prevalence of HSV2, the lifelong nature of the infection, limitations in diagnosis and treatment, and the bi-directional interaction with HIV, suggest that HSV2 may be responsible for a substantial proportion of HIV infections in some African countries. Furthermore, although the effect of bacterial sexually transmitted diseases (STD) may decline in mature HIV epidemics, as HIV infection spreads beyond subgroups of the population with high STD rates [25], HSV2 may become more important for several reasons. First, the high prevalence of HSV2 means that a large proportion of the sexually active population is affected, and infection is not restricted to subgroups with high rates of partner change. Second, as the number of persons with advanced stages of HIV infection increases, the incidence and severity of HSV2 clinical episodes among such individuals is likely to increase. Third, this in turn may lead to an overall increase in HSV2 incidence among both HIV-negative and HIV-positive subjects [26,27].

Despite the potential importance of HSV2 for HIV transmission, few studies have measured the association between these infections in the general population. Nearly all of these studies have been cross-sectional in design, and while strong associations have been reported it is difficult to establish whether HSV2 infection preceded HIV or vice versa. We are aware of only one previous prospective study in Africa that has examined the association between HSV2 infection and HIV seroconversion. In this study among male factory workers in Zimbabwe [20], adjusted rate ratios of HIV seroconversion were 3.5 in men who were HSV2-positive at baseline, and 6.7 in men who seroconverted to HSV2 during follow-up, compared to men who remained HSV2 negative.

The follow-up of a large population-based cohort in rural Tanzania, in the context of a randomized controlled trial of improved STD treatment services, afforded the opportunity to examine the relative importance of HSV2 infection in the transmission of HIV in this population. The objectives of this study were to investigate the association between HIV incidence and HSV2 infection at baseline or occurring during the 2-year follow-up period; to compare these associations between men and women, between younger and older age-groups, and between the study arms of the intervention trial; and to assess the proportion of new HIV infections that could be attributed to HSV2 infection in this population.


Study design

A randomized controlled trial was conducted in 12 rural communities in Mwanza Region, Tanzania between 1991 and 1994. The objective was to evaluate the impact of improved STD management on the incidence of HIV infection. The intervention was integrated into routine government health services, and consisted mainly of training and supervision of primary health care workers to deliver syndromic STD treatment, a reliable supply of STD drugs and health education to improve treatment-seeking behaviour [28]. Communities were arranged into six pairs on the basis of geographical characteristics, and one community in each pair was randomly allocated to receive the intervention. The evaluation was performed in a cohort of 12 537 individuals aged 15–54 years (approximately 1000 persons from each community). HIV incidence was reported to be 40% lower in the intervention communities [28]. The present analysis corresponds to an unmatched case–control study nested within the trial cohort.

Study population

Characteristics of the study population have been described previously [28]. In brief, the region is located on the southern shores of Lake Victoria. Most of its two million inhabitants live in rural villages where subsistence farming is the main activity [29]. The Sukuma are the predominant ethnic group, but other smaller ethnic groups are also present. Only 19% of men are circumcised [11].

Recruitment and follow-up of the cohort

Briefly, 20–30 clusters of households were defined in each community and between seven and nine of these were randomly sampled. All adults aged 15–54 years living in these households were eligible for recruitment to the cohort. Overall, 85% of eligible adults were recruited [30].

To compare HIV incidence in the six intervention and six comparison communities, two consecutive serological surveys were performed. Of those recruited to the cohort, 8845 (71%) were seen again at follow-up 2 years later (Fig. 1). Reasons for loss to follow-up were: 20% had moved away or were temporarily absent, 2% had died and 8% did not participate for other reasons (refusals, inaccessibility due to heavy rains, or reason not recorded) [28].

Fig. 1.
Fig. 1.:
Study profile, Mwanza Region, Tanzania. SBS, Sexual behaviour survey.

Selection of cases and controls

For the purposes of the present study, cases of HIV seroconversion and appropriately chosen controls were tested for HSV2 serology at baseline and after 2 years of follow-up.

A total of 130 HIV seroconversions were observed between baseline and follow-up in the trial cohort and these subjects were eligible as cases for the present study. HSV2 results were unavailable for three subjects, leaving 127 cases for analysis (Fig. 1).

To estimate HSV2 prevalence and incidence by age and sex, and to provide controls for the case–control analysis, a subcohort of 671 was selected randomly from the 8845 cohort members seen at both baseline and follow-up. The sampling of the subcohort was weighted towards the younger age groups, because most subjects are already HSV2 positive at older ages [18]. The sampling fractions by age group were 340 out of 1413 (24%) of those aged 15–19 years, 157 out of 1626 (10%) of those aged 20–24 years, and 174 out of 5806 (3%) of those aged 25 years and over.

The random subcohort included 10 HIV seroconverters (who were among the 130 cases), and 13 subjects who were already HIV-positive at baseline. After excluding these 23 subjects, there were 648 persistently HIV-negative controls. HSV2 results were unavailable for 12 subjects, leaving 636 controls for analysis.

Sexual behaviour survey

Basic sociodemographic data were available for all members of the trial cohort. To obtain more detailed information on behavioural risk factors, a sexual behaviour survey (SBS) was conducted in a subset of subjects 2–7 months after the follow-up survey.

All HIV seroconverters (cases) were eligible as participants, but in the event 11 out of 127 cases were not sought because of delays in confirming HIV serological tests. Of the 116 cases sought, 94 were successfully recruited to the SBS, and after excluding five individuals for whom there were doubts about correct identity, detailed sexual behaviour data were available for 89 cases (Fig. 1).

A random sample of 349 out of 636 controls was sought for the SBS, and 293 were successfully recruited. After excluding 12 controls for whom there were doubts about correct identity, detailed sexual behaviour data were available for 281 controls (Fig. 1).

Face-to-face interviews were conducted in private using structured pre-coded questionnaires, which were developed in English, translated into Kiswahili and back-translated into English [2,4].

Laboratory methods

HIV serology was determined using enzyme-linked immunosorbent assay (ELISA; Vironostika HIV MIXT Microelisa, Organon Technika, Boxtel, the Netherlands). All positive samples were subjected to a confirmatory test with a methodologically independent ELISA (Wellcozyme HIV 1+2 GACELISA, Murex Diagnostics, Dartford, Kent, UK). To evaluate discordant or indeterminate samples, a Western blot was performed (HIV-1 Westernblot, Epitope, Beaverton, Oregon, USA) [4].

Type-specific antibodies to HSV2 were detected using a monoclonal blocking ELISA, validated for seroepidemiological studies in Africa [31]. In brief, Maxisorp plates were coated with HSV2 antigen diluted in phosphate-buffered saline (PBS). Plates were washed five times with PBS, blocked and washed a further five times with PBS. Test and control sera were diluted 1 : 2 in 0.3% non-fat milk/PBS/0.05% Tween-20. Sera were incubated for 3 h with shaking at 37°C, after which plates were washed five times with PBS 0.05% Tween-20 and incubated for 1 h at 37°C with 100 μl type-specific monoclonal antibody. Plates were then washed three times with PBS 0.05% Tween-20, and bound AP-1 was detected with anti-mouse horseradish peroxidase.

Statistical analysis

Logistic regression was used to examine the association between HSV2 and HIV, with HIV seroconversion as response variable, and HSV2 status as exposure variable. HSV2 status was analysed in three categories: those remaining HSV2 negative at the end of follow-up (reference category); those already HSV2 positive at baseline; and those who seroconverted to HSV2 during the 2-year follow-up period. Tests for trend over these ordered categories were obtained by fitting models with HSV2 entered as a numerical variable (coded 0, 1, 2).

All analyses were adjusted for age group at baseline (15–19, 20–24, 25–34, 35–44, 45–54 years) and community of residence. Adjustment for age group was necessary to ensure valid estimates of odds ratios (OR) given the age-weighted sampling of controls. Further adjustment was made for other confounding factors if the OR for HSV2 was changed by 20% or more.

Subgroup analyses were carried out on three variables selected a priori. Separate analyses were undertaken in men and women, as the effects of HSV2 and other risk factors may vary between the sexes. Subgroup analyses were also conducted in subjects younger than and older than 25 years, as any effect of prevalent HSV2 infection may differ according to time from primary infection; and in the intervention and comparison arms of the intervention trial, as the relative effect of HSV2 may have changed if the STD treatment intervention led to reductions in other STD.

Adjusted OR were used to estimate the population attributable fraction (PAF) of HIV seroconversions due to HSV2 separately for men and women [32]. PAF estimates were also obtained by age group (younger and older than 25 years) and trial arm.

All statistical analyses were performed using Stata 6.0 (Stata, University Drive East, College Station, USA). Statistical significance was assessed using the likelihood ratio test (LRT).


Study population

The age and sex distribution of cases and controls is shown in Table 1. The controls were younger on average than the cases, reflecting the age-weighted sampling of controls. All analyses of the association between HIV and HSV2 were therefore adjusted for age.

Table 1
Table 1:
Age and sex distribution of cases and controls, and participation in the sexual behaviour survey (SBS).

Among those selected for the SBS overall participation rates were 77% (89/116) for cases and 81% (281/349) for controls (Table 1). Participants and non-participants were similar with respect to age, sex and most other socio-demographic factors (data not shown). Among men, non-participants were less likely to be from the Sukumu ethnic group than participants (56% versus 76%) and more likely to have lived away from the village during the past year (44% versus 23%). Among women, non-participants were more educated than those who participated in the survey (4+ years of schooling, 70% versus 54%), and more likely to be Christian (87% versus 75%).

Prevalence and incidence of HSV2 infection

Of the 671 subjects selected for the random subcohort (Fig. 1) HSV2 test results were available for 665 (99.1%) at baseline and 660 (98.4%) at follow-up. HSV2 seroprevalence at baseline, and seroincidence during the 2-year follow-up period among those initially seronegative, are shown by age and sex in Table 2.

Table 2
Table 2:
Seroprevalence and seroincidence ofHerpes simplex virus type-2 by age and sex in a random sample of adults aged 15–54 years in Mwanza Region, Tanzania.

HSV2 seroprevalence increased rapidly with age, reaching a plateau of around 50–60% among men and 75% in women aged 30 years and over. Prevalence in women aged 15–19 years was nearly three times that in men of the same age, and prevalence remained higher in women in every age-group.

During the 2-year follow-up period, HSV2 seroconversion was recorded in 17.5% of 206 women negative at baseline, and in 11.3% of 221 men negative at baseline. Seroincidence was around 5% per year in men of all ages and in women aged 20 years and over, but was much higher at around 10% per year in women aged 15–19 years.

Association between HIV seroconversion and HSV2 infection

Among men, 86% (60/70) of HIV seroconversions took place in subjects with HSV2 infection (Table 3). There was a highly significant association between HIV sercoconversion and HSV2 status, which persisted after adjustment for age and community of residence (test for trend, P < 0.001). Compared with those remaining HSV2 negative at follow-up, HIV incidence was substantially higher in men who were HSV2 positive at baseline (adjusted OR, 5.78), and even higher in those who seroconverted to HSV2 during follow-up (adjusted OR, 13.2).

Table 3
Table 3:
Association between HIV seroconversion andHerpes simplex virus type-2 (HSV2) infection among men and women in Mwanza Region, Tanzania.

When the analysis was adjusted for other factors, religion was the only variable that appeared to have a substantial confounding effect. After adjusting for age, community and religion, the association between HIV and HSV2 became even stronger (Table 3), with adjusted OR of 6.12 in men who were HSV2 positive at baseline, and 16.8 in HSV2 seroconverters. Further adjustment for other factors that were measured for all cohort members, such as education, occupation, ethnic group, living away from the village, Treponema pallidum haemagglutination assay (TPHA) status and history of STD, had little effect on the association (data not shown).

Among women, 74% (42/57) of HIV seroconversions occurred in HSV2-infected subjects. However, the association between HIV seroconversion and HSV2 infection was weaker than in men (Table 3). Compared with those remaining HSV2 negative at follow-up, HIV incidence was only slightly raised in women who were HSV2 positive at baseline (adjusted OR, 1.32), and moderately raised in those who seroconverted to HSV2 (adjusted OR, 2.36). The association was not statistically significant (P = 0.14). Adjustment for religion (Table 3), and other factors that were measured for all cohort members (data not shown), had little effect on the association.

When analysis was restricted to the subsample of individuals who participated in the SBS, adjustment for behavioural factors was limited by the small sample size and wide confidence intervals reflecting the statistical colinearity between variables. However, there was no evidence of appreciable confounding by these variables in either sex. For example, adjustment for reported number of sexual partners during the follow-up period had little effect on the association between HIV and HSV2 (data not shown).

Age-specific effects

Among men, the effect of HSV2 infection varied significantly depending on the age of the subject (LRT for interaction, P = 0.04, adjusted for age and community), and age-specific associations are shown in Table 4. HSV2 infection at baseline was associated with a much greater increase in HIV risk in young men aged 15–24 years than in older men (adjusted OR of 14.4 and 5.00, respectively). HSV2 seroconversion showed a strong association with HIV in both age groups, although the relative risk was larger in older men (adjusted OR of 9.23 and 45.2, respectively).

Table 4
Table 4:
Association between HIV seroconversion andHerpes simplex virus type-2 (HSV2) infection in men and women according to age group and arm of intervention trial.

Among women, there was no significant variation in HSV2 effects between age groups (LRT for interaction, P = 0.57, adjusted for age and community).

Analysis by trial arm

Table 4 also shows associations between HIV seroconversion and HSV2 infection separately for the intervention and comparison arms of the intervention trial. Among men, the association was strongly significant in both arms of the trial, while among women a stronger association was seen in the intervention arm than the comparison arm (adjusted OR, 1.69 and 1.14, respectively for those HSV2 positive at baseline; 4.08 and 1.69, respectively for HSV2 seroconverters). However, the differences between arms were not statistically significant (LRT for interaction, P = 0.91 in men and P = 0.64 in women, adjusted for age and community).

Fraction of HIV seroconversions attributable to HSV2

The proportion of HIV seroconverters that was HSV2 positive and the adjusted OR for the association between HIV and HSV2 were used to estimate the PAF, the fraction of HIV seroconversions that could be attributed to HSV2. PAF estimates are given for men and women in Table 5. Estimates are also given for younger and older age groups, and for the intervention and comparison arms of the trial.

Table 5
Table 5:
Fraction of HIV seroconversions attributable toHerpes simplex virus type-2 (HSV2) infection (PAF)a by age-group and arm of intervention trial.

Among men, 45% of HIV seroconversions were attributable to prevalent HSV2 infection (HSV2 positive at baseline), and 29% to incident HSV2 infection (HSV2 seroconverters), giving an overall PAF of 74%. PAF were generally similar in older and younger age groups, and in the two arms of the trial.

A much lower PAF was observed among women. Only 15% of HIV seroconversions were attributable to prevalent HSV2 infection, and 7% to incident HSV2 infection, giving an overall PAF of 22%. The PAF was somewhat higher in older women (41%) than in younger women (12%). There was also a difference between the trial arms, with PAF of 36% in the intervention arm and 12% in the comparison arm. However, these differences were not statistically significant.


This is the first longitudinal study to report the association between HIV incidence and HSV2 infection in a random sample of adults drawn from the general population. Strengths of the study include the representative sample, the large sample size, the longitudinal design (allowing estimation of the effects of both prevalent and incident HSV2 infections) and the availability of data on potential confounding factors. Nevertheless there are a number of potential sources of error that require careful consideration.

Sources of error

Cases and controls were drawn from the same population and participation rates in the baseline survey were high (85%) [30]. Of those recruited to the trial cohort, 71% were successfully followed up after 2 years. Participation rates among those selected for the SBS were also high (80%). The main reasons for non-participation and loss to follow-up were temporary absence or moving permanently away from the village. Migration and travel have been shown to be associated with higher HIV risk [29,33] in East African studies. While this may have resulted in underestimation of HSV2 prevalence and incidence in this cohort, it is unlikely to have appreciably influenced the measured association between HIV and HSV2.

The HIV testing algorithm used in this study is highly sensitive and specific, and there are likely to be very few false-positive or false-negative HIV results. The serological assay for HSV2 has been validated for use in epidemiological studies in Africa, and shown to have 93% sensitivity and 91% specificity [31]. This will have led to some misclassification of HSV2 infection status, which should be non-differential as HSV2 testing was carried out blind to HIV status and other variables, leading to under-estimation of the true association.

HIV and HSV2 share a similar mode of transmission, so that an epidemiological association may reflect common risk factors for the acquisition of the two infections. Adequate adjustment for confounders is therefore important. We collected data on a wide range of socio-demographic and behavioural variables. Although questionnaires were carefully designed, and interviewers received intensive training and supervision, our data on potential confounders are clearly subject to error, particularly those concerning sensitive variables such as reported number of sexual partners [34]. Errors in such variables may have led to residual confounding, implying either under-estimation or over-estimation of the association of HIV and HSV2. A further limitation was that data on sexual behaviour were collected only from a sample of the cohort, and the smaller sample size restricted the precision of adjusted estimates. However, there was little evidence of confounding by measured risk factors with the exception of religion, adjustment for which led to strengthening of the association between HIV and HSV2 in men. The apparent lack of confounding by sexual behaviour may reflect the differing epidemiology of the two infections. HSV2 has a high basic reproduction number, so that infection is not restricted to subgroups of the population with high rates of sexual partner change [13,35–37].

In summary, while several possible sources of error have been identified, it seems unlikely that these could account for the strong associations observed in this study, particularly among men.

Association and causation

Assuming that the observed association between HIV and HSV2 is not due to bias or confounding, it is likely to reflect a causal association between the two infections. The direction of causation is difficult to establish from cross-sectional studies, but the longitudinal design of the current study partly resolves this problem. Thus, the increased HIV incidence seen in men who were already HSV2 positive at baseline supports the hypothesis that HSV2 infection enhanced susceptibility to HIV acquisition.

The association between HIV and HSV2 seroconversion is more difficult to interpret, as we cannot establish which infection occurred first. The longitudinal study conducted previously among factory workers in Zimbabwe [20] found an increased rate of HSV2 seroconversion among men who were already HIV-positive at baseline, possibly reflecting increased susceptibility to HSV2 due to HIV-related immunosuppression. In the present study, it seems less likely that the level of immunosuppression shortly after HIV infection would be sufficient to enhance HSV2 acquisition, but we cannot rule this out. A more plausible explanation, however, is that the severe genital lesions often associated with primary HSV2 infection lead to greatly increased risk of HIV acquisition. A further possible mechanism is that dually-infected sexual partners of our study participants were more likely to transmit HIV due to an effect of HSV2 on HIV infectiousness. In this case the observed association between HIV and HSV2 seroconversion would still reflect an effect of HSV2 in enhancing HIV transmission, but through an effect on infectiousness rather than susceptibility.

An alternative explanation for both associations (with prevalent or incident HSV2) is that HIV-infected sexual partners of study participants were more likely to transmit HSV2, leading to an excess of participants acquiring both infections. Such an effect is possible, since HSV2 shedding has been shown to be increased in HIV-positive subjects [38–40]. The transmission probability of HSV2 per single sexual contact in the absence of HIV is very much higher than the transmission probability of HIV in the absence of HSV2, and so it seems plausible that the relative effect of HSV2 on HIV transmission may be stronger than the effect of HIV on HSV2 transmission. However, we cannot estimate the extent to which the observed association is due to these two alternative mechanisms. The measured associations may therefore over-estimate somewhat the effects of HSV2 on HIV infection in this population.

Association between HIV and HSV2 in men and women

The most striking finding of this study was the extremely strong association between HIV seroconversion and HSV2 infection among men. With adjusted odds ratios of up to 17 in HSV2 seroconverters, it is highly unlikely that such effects were due to confounding or other sources of error. The most likely explanation is that the probability of HIV transmission is greatly increased in the presence of herpetic genital lesions due to HSV2.

The much greater risk of HIV in those seroconverting to HSV2 during the study period (OR, 16.8) compared to those already HSV2 positive at baseline (OR, 6.12) is biologically plausible. HSV2 seroconversion indicates primary infection, which is likely to be associated with more prolonged and severe genital lesions than recurrent herpes episodes [41–43]. Some subjects who were HSV2 positive at baseline may have been infected many years previously, and the frequency of clinical episodes is thought to decrease over time. Similar findings were observed in the only previous longitudinal study conducted in Zimbabwe [20].

A similar trend was seen among women, but the association was much weaker. There are several possible explanations for this sex difference. First, the difference may be due to random error, but this seems unlikely as a test for interaction between sex and HSV2 was highly significant (P = 0.01). Second, the more rapid increase in HSV2 positivity with age suggests that women tend to be infected at a younger age than men. This would imply that HSV2 prevalent cases in women would more often represent long-established infections, with a lower frequency and severity of clinical episodes of herpes. This would help to explain the weaker association with prevalent HSV2 among women, but not the weak association with HSV2 seroconversion. Finally, the weaker effect of HSV2 in women may reflect differences in the underlying transmission dynamics of HIV infection in men and women in this population. Because the transmission probability of HIV from men to women seems to be higher than from women to men [44–46], the baseline probability of infection in women exposed to HIV-positive partners was probably several times higher than in men. Thus, even if the absolute effects of HSV2 were similar, lower relative risks would be observed.

Age-specific effects

A much stronger association was observed between HIV seroconversion and HSV2 positivity at baseline among younger men aged 15–24 years than among those aged 25 years and over. This may again reflect the more recent acquisition of HSV2 infection among younger HSV2-positive men, implying a greater frequency and severity of clinical herpes [41–43]. We have no explanation for the stronger association with HSV2 seroconversion in older men. Interestingly, this was also observed among women, although there was no significant interaction between the effects of age group and HSV2 indicating that the observed differences in OR may have been due to chance.

Analysis by trial arm

The present study was nested within a trial of improved STD treatment services. The intervention involved syndromic treatment for treatable STD, so that cases of genital ulcer syndrome were treated with drugs effective against chancroid and syphilis, the main treatable causes of genital ulcers in this population. No treatment was available for genital herpes.

We expected to find a stronger association between HIV seroconversion and HSV2 in the intervention arm of the trial. The intervention was designed to reduce the duration of treatable STD, and thus their enhancing effect on HIV transmission, so HSV2 was expected to be of relatively greater importance in this arm. There was little evidence of a stronger association in the intervention arm among men. In women, the association of HIV seroconversion with baseline HSV2 positivity and HSV2 seroconversion was stronger in the intervention arm, as expected. However sample sizes for this subgroup analysis were small, and the observed interaction may have occurred by chance.

Fraction of HIV seroconversions attributable to HSV2

Our estimates of attributable fractions assume that the observed association between HIV and HSV2 reflects a causal relationship in which HSV2 enhances HIV transmission, that confounders have been adequately controlled, and that exposure and outcome are accurately measured. These estimates should therefore be interpreted with caution, but are consistent with the hypothesis that HSV2 infection was responsible for a large proportion of new HIV infections. Although confidence intervals were wide, we estimated that 74% of HIV seroconversions in men were attributable to HSV2 but only 22% of those in women. Possible reasons for the weaker effect of HSV2 in women have already been discussed.

While there was no evidence that the attributable fraction varied between trial arms in men, there was a marked difference in women with PAF of only 12% in the comparison arm, but 36% in the intervention arm. As treatment for curable STD is strengthened, with a reduction in the duration and prevalence of these infections, it seems likely that genital herpes will become relatively more important in enhancing the spread of HIV.

Implications for HIV control

Our results add to a growing body of evidence that HSV2 may play a major role in HIV transmission in Africa. Cross-sectional studies have shown strong associations between HIV and HSV2, but these are difficult to interpret as the time sequence of infections cannot be determined, and it is difficult to adjust adequately for confounding by sexual behaviour because the period of exposure is ill-defined. The present study, together with the previous occupational study in Zimbabwe, partly avoid these limitations through their longitudinal design. Both studies show strong associations, although it will be of interest to see whether the weaker association observed among women in Mwanza is replicated in other populations.

In a recent ecological study in four cities in Africa, designed to identify factors that might explain why two cities have experienced explosive HIV epidemics while in the other two HIV prevalence has remained relatively low, few differences were apparent in patterns of sexual behaviour [47]. However, there were two biological factors that showed clear differences. First, male circumcision was almost universal in the low prevalence cities but uncommon in the high prevalence cities. Second, infections associated with genital ulcer disease, and in particular HSV2 infection, were more common in the high prevalence cities.

This ecological study, together with the strong associations observed in epidemiological studies, and supporting clinical studies identifying plausible biological pathways, strongly support a causal role for HSV2 in enhancing HIV infection. Our attributable fraction estimates suggest that a large proportion of new HIV infections in some populations may be attributable to HSV2, implying that HSV2 control may be an effective strategy for HIV prevention in populations where HSV2 is prevalent. What options are available for control?

Episodic or suppressive treatment of HSV2 may help to reduce the frequency, duration or severity of herpetic episodes, and hence reduce the effect of HSV2 on HIV transmission. However, wide-scale provision of treatment is currently complicated by the high cost of drugs, as well as logistical difficulties in identifying those requiring treatment in large populations, although it might be feasible to provide treatment to high-risk groups.

In terms of primary prevention, measures to promote safer sexual behaviour should impact on HSV2 incidence, but consequent changes in HSV2 prevalence may take many years given the large reservoir of infected individuals. It has been suggested that safe services for male circumcision might be effective in reducing HIV incidence in populations where circumcision is uncommon. It is possible that HSV2 rates are also lower in circumcised men [48], and this may help to explain the geographical variations in HSV2 prevalence observed in the four-cities study [47]. Thus increased rates of male circumcision could reduce HIV spread both directly, and indirectly by reducing the incidence of genital herpes and possibly other ulcerative STD.

Considering the limitations of other approaches to HSV2 control, an effective HSV2 vaccine, designed either to prevent primary infection or to reduce clinical expression in HSV2-infected subjects, could play a valuable role in preventing HIV spread. Given the public health importance of the HIV epidemic, and the evidence suggesting that a large proportion of new HIV infections may be attributable to the enhancing effect of HSV2, the development of such a vaccine should be regarded as a high global health priority.


We thank the Principal Secretary, Ministry of Health, the manager of the National AIDS Control Programme, and the Director General of the National Institute for Medical Research, Tanzania for permission to carry out and publish the results of this study. We also thank the Regional Medical Officer, Mwanza, the Director of the National Institute for Medical Research, Mwanza, the Director of the African Medical and Research Foundation, Mwanza and regional, district, ward and community leaders for the support provided. We are particularly grateful to the study participants and to the field research team. We thank the referees for their helpful comments and suggestions.


1. WHO/UNAIDS. AIDS epidemic update: December 2000. Geneva: UNAIDS; 2000.
2. Munguti K, Grosskurth H, Newell J. et al. Patterns of sexual behaviour in a rural population in Northwestern Tanzania. Soc Sci Med 1997, 44: 1553–1561.
3. Buvé A, Caraël M, Hayes R, Robinson NJ. Variations in HIV prevalence between urban areas in sub-Saharan Africa: do we understand them? AIDS 1995, 9 (suppl A): S103–S109.
4. Quigley M, Munguti K, Grosskurth H. et al. Sexual behaviour patterns and other risk factors for HIV infection in rural Tanzania: a case-control study. AIDS 1997, 11: 237–248.
5. Rakwar J, Lavreys L, Thompson ML. et al. Cofactors for the acquisition of HIV-1 among heterosexual men: prospective cohort study of trucking company workers in Kenya. AIDS 1999, 13: 607–614.
6. Weiss HA, Quigley MA, Hayes R. Male circumcision and risk of HIV infection in sub-Saharan Africa: a systematic review and meta-analysis. AIDS 2000, 14: 2361–2370.
7. Lavreys L, Rakwar JP, Thompson ML. et al. Effect of circumcision on incidence of human immunodeficiency virus type 1 and other sexually transmitted diseases. J Infect Dis 1999, 180: 330–336.
8. Kelly R, Kiwanuka N, Wawer MJ. et al. Age of male circumcision and risk of prevalent HIV infection in rural Uganda. AIDS 1999, 13: 399–405.
9. Moses S, Bradley J, Nagelkerke N, Ronald A, Ndinya-Achola J, Plummer F. Geographical patterns of male circumcision practices in Africa: association with HIV prevalence. Int J Epidemiol 1990, 19: 693–697.
10. Serwadda D, Wawer MJ, Musgrave SD, Sewankambo NK, Kaplan JE, Gray RH. HIV risk factors in three geographical strata of rural Rakai District. AIDS 1992, 6: 983–989.
11. Urassa M, Todd J, Boerma JT, Hayes R, Isingo R. Male circumcision and susceptibility to HIV infection among men in Tanzania. AIDS 1997, 11: 73–80.
12. O′Farrell N. Increasing prevalence of genital herpes in developing countries: implications for heterosexual HIV transmission and STI control programmes. Sex Transm Infect 1999, 75: 377–384.
13. Corey L, Handsfield HH. Genital herpes and public health: addressing a global problem. JAMA 2000, 283: 791–794.
14. Brugha R, Keersmaekers K, Renton A, Meheus A. Genital herpes infection: a review. Int J Epidemiol 1997, 26: 698–709.
15. Shomogyi M, Wald A, Corey L. Herpes simplex virus-2 infection.An emerging disease? Infect Dis Clin North Am 1998, 12: 47–61.
16. Siegel D, Golden E, Washington AE. et al. Prevalence and correlates of herpes-simplex infections.The population-based AIDS in multi-ethnic neighbourhoods study. JAMA 1992, 268: 1702–1708.
17. Kamali A, Nunn AJ, Mulder DW, Van Dyck E, Dobbins JG, Whitworth JA. Seroprevalence and incidence of genital ulcer infection in a rural Ugandan population. Sex Transm Infect 1999, 75: 98–102.
18. Obasi A, Mosha F, Quigley M. et al. Antibody to herpes simplex virus type 2 as a marker of sexual risk behaviour in rural Tanzania. J Infect Dis 1999, 179: 16–24.
19. Wawer MJ, Sewankambo NK, Serwadda D. et al. Control of sexually transmitted diseases for AIDS prevention in Uganda: a randomised community trial. Lancet 1999, 353: 525–535.
20. 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 1999, 180: 1459–1465.
21. Gwanzura L, McFarland W, Alexander D'A, Burke RL, Katzenstein D. Association between human immunodeficiency virus and herpes simplex virus type 2 seropositivity among male factory workers in Zimbabwe. J Infect Dis 1998, 177: 481–484.
22. Chen CY, Ballard RC, Beck Sague CM. et al. Human immunodeficiency virus infection and genital ulcer disease in South Africa: The herpetic connection. Sex Transm Dis 2000, 27: 21–29.
23. Bagdades EK, Pillay D, Squire SB, O'Neil C, Johnson MA, Griffiths PD. Relationship between herpes simplex virus ulceration and CD4+ cell counts in patients with HIV infection. AIDS 1992, 6: 1317–1320.
24. Schacker T, Zeh J, Hu HL, Hill E, Corey L. Frequency of symptomatic and asymptomatic herpes simplex virus type 2 reactions among human immunodeficiency virus-infected men. J Infect Dis 1998, 178: 1616–1622.
25. Robinson NJ, Mulder DW, Auvert B, Hayes RJ. Proportion of HIV infections attributable to other sexually transmitted diseases in a rural Ugandan population: simulation model estimates. Int J Epidemiol 1997, 26: 180–189.
26. Grosskurth H, Gray R, Hayes R, Mabey D, Wawer M. Control of sexually transmitted diseases for HIV-1 prevention: understanding the implications of the Mwanza and Rakai trials. Lancet 2000, 355: 1981–1987.
27. Fleming DT, Wasserheit JN. From epidemiological synergy to public health policy and practice: the contribution of other sexually transmitted diseases to sexual transmission of HIV infection. Sex Transm Infect 1999, 75: 3–17.
28. Grosskurth H, Mosha F, Todd J. et al. Impact of improved treatment of sexually transmitted diseases on HIV infection in rural Tanzania: randomised controlled trial. Lancet 1995, 346: 530–536.
29. Barongo LR, Borgdorff MW, Mosha FF. et al. The epidemiology of HIV-1 infection in urban areas, roadside settlements and rural villages in Mwanza Region, Tanzania. AIDS 1992, 6: 1521–1528.
30. Grosskurth H, Mosha F, Todd J. et al. A community trial of the impact of improved sexually transmitted disease treatment on the HIV epidemic in rural Tanzania: 2 Baseline survey results. AIDS 1995, 9: 927–934.
31. Gopal R, Gibbs T, Slomka MJ. et al. A monoclonal blocking EIA for herpes simplex virus type 2 antibody: validation for seroepidemiological studies in Africa. J Virol Methods 2000, 87: 71–80.
32. Brady AR. Adjusted population attributable fractions from logistic regression. Stata Technical Bulletin 1998, 42: 137–143.
33. Nunn AJ, Kengeya Kayondo JF, Malamba SS, Seeley JA, Mulder DW. Risk factors for HIV-1 infection in adults in a rural Ugandan community: a population study. AIDS 1994, 8: 81–86.
34. Catania JA, Gibson DR, Chitwood DD, Gates TJ. Methodological problems in AIDS behavioral research: influences on measurement error and participation bias in studies of sexual behavior. Psychol Bull 1990, 108: 339–362.
35. Bryson Y, Dillon M, Bernstein DI, Radolf J, Zakowski P, Garatty E. Risk of acquisition of genital herpes simplex virus type 2 in sex partners of persons with genital herpes: a prospective couple study. J Infect Dis 1993, 167: 942–946.
36. Mertz GJ, Benedetti J, Ashley RA, Selke S, Corey L. Risk factors for the sexual transmission of genital herpes. Ann Intern Med 1992, 116: 197–202.
37. Nahmias AJ, Lee FK, Beckman-Nahmias S. Seroepidemiological patterns of herpes simplex virus infection in the world. Scand J Infect Dis Suppl 1990, 69: 19–36.
38. Augenbraun M, Feldman J, Chirgwin K. et al. Increased genital shedding of herpes simplex virus type 2 in HIV-seropositive women. Ann Intern Med 1995, 123: 845–847.
39. 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 2000, 182: 1090–1096.
40. Schacker T, Hu H, Koelle DM. et al. Famciclovir for the suppression of symptomatic and asymptomatic herpes simplex virus reactivation in HIV-infected persons. Ann Intern Med 1998, 128: 21–28.
41. Benedetti JK, Zeh J, Corey L. Clinical reactivation of genital herpes simplex virus infection decreases in frequency over time. Ann Intern Med 1999, 131: 14–20.
42. Diamond C, Selke S, Ashley R, Benedetti J, Corey L. Clinical course of patients with serologic evidence of recurrent genital herpes presenting with signs and symptoms of first episode disease. Sex Transm Dis 1999, 26: 221–225.
43. Koelle DM, Benedetti K, Langenberg A, Corey L. Asymptomatic reactivation of herpes simplex virus in women after the first episode of genital herpes. Ann Intern Med 1992, 116: 433–437.
44. Carpenter LM, Kamali A, Ruberantwari A, Malamba SS, Whitworth JA. Rates of HIV-1 transmission within marriage in rural Uganda in relation to the HIV sero-status of the partners. AIDS 1999, 13: 1083–1089.
45. Nicolosi A, Correa Leite ML, Musicco M, Arici C, Gavazzeni G, Lazzarin A. The efficiency of male-to-female and female-to-male sexual transmission of the human immunodeficiency virus: A study of 730 stable couples. Epidemiology 1994, 5: 570–575.
46. European Study Group on Heterosexual Transmission of HIV. Comparison of female to male and male to female transmission of HIV in 563 stable couples. Br Med J 1992, 304: 809–813.
47. Buvé A, Caraël M, Hayes RJ. et al. The multicentre study on factors determining the differential spread of HIV in four African cities: summary and conclusions. AIDS 2001, 15 (suppl 4): S127–S131.
48. Weiss HA, Buvé 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.

HIV; Herpes simplex virus type 2; epidemiology; case–control study; risk factors; Africa

© 2002 Lippincott Williams & Wilkins, Inc.