Herpes simplex virus type 2 (HSV-2) is one of the most common sexually transmitted infections (STIs) worldwide and the leading cause of genital ulcer disease [Reference Fanfair1, Reference Looker, Garnett and Schmid2]. HSV-2 is transmitted through symptomatic and asymptomatic viral shedding [Reference Gupta, Warren and Wald3]. The synergistic relationship between HIV-1 and HSV-2 transmission has been well documented, with enhanced transmission and/or acquisition of HIV-1 or HSV-2 in the presence of the other virus [Reference Corey4–Reference Mahiane10].
Heterosexual contact has emerged recently as the primary mode of HIV transmission in China, accounting for 52·2% of estimated new infections in 2011 . It is noteworthy that an increased proportion of new infections occurred in couples engaged in stable, heterosexual HIV-1-discordant relationships (i.e. one partner is HIV-1 infected and the other is not) [Reference He12, Reference Jia13]. High HSV-2 prevalence is observed in one or both members of HIV-1-discordant couples in many countries including China [Reference Muiru14–Reference Zhu16]. Prevention of HSV-2 infection, as part of HIV-1 prevention programmes, may be an important component in controlling the risk of HIV-1 transmission in these couples [Reference Mbopi-Kéou9]. However, few data exist on the prevalence and risk factors of HSV-2 infection in this population in China.
In HIV-1-discordant couples, HSV-2 can be acquired either from within or outside the couple's relationship. To inform design of interventions it is important to separately identify risk factors that independently affect HSV-2 infection from outside the partnership or HSV-2 transmission within the couple. Most HSV-2 discordance arise due to HSV-2 infection of one partner from sexual activity outside the partnership, but HSV-2-positive concordance may reflect acquisition outside the partnership or HSV-2 transmission within the couple [Reference Muiru14, Reference Mujugira17]. Therefore, we separately assessed risk factors that were independently associated with HSV-2 discordance compared to HSV-2-negative concordance as well as risk factor that were independently associated with HSV-2-positive concordance compared to HSV-2 discordance in a sample of HIV-1-discordant couples.
This study used the baseline data from a prospective, longitudinal cohort study of HIV-1-discordant couples conducted from June 2009 to March 2011 in Dehong prefecture of Yunnan province, which is a region bordering Myanmar and is one of the areas most affected by HIV in China [Reference Jia18], as described previously [Reference He12]. All HIV-1-discordant couples in Dehong who were registered with the Chinese National Information System for AIDS Prevention and Control (CNISAPC) were invited to participate in this study [Reference He12]. Each study participant was administered a face-to-face confidential questionnaire interview by a trained public health worker in a private setting, usually the participant's home. The questionnaire captured data on sociodemographic characteristics and sexual behaviours.
The study protocol, including study design, subject recruitment, consent procedure, etc., was reviewed and approved by the Institutional Review Board of the Chinese National Centre for AIDS/STD Control and Prevention, Chinese Centre for Disease Control and Prevention. Written informed consent was obtained from all participants, with one copy given to the study subject. Each participant was given compensation of CNY60 (~US$10) for their time. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
HIV-1, syphilis and HSV-2
HIV testing was performed using a commercially available enzyme-linked immunosorbent assay (ELISA) kit (Kehua Biotech, China). All samples screened positive for HIV-1 were confirmed by Western blot (Genelabs Diagnostics, Singapore). Participants with plasma positive for syphilis in both the rapid plasma regain test and Treponema pallidum haemagglutination assay were determined to be currently infected with syphilis. HSV-2 IgG antibody testing was performed using the Captia anti-HSV-2 IgG ELISA kit (Trinity Biotech, Ireland).
All statistical analyses were performed using SAS software v. 9·11 (SAS Institute Inc., USA). First, descriptive analyses were conducted to elucidate the sociodemographic characteristics and sexual behaviours of the study couples by gender and HIV-1 serostatus. Differences by gender and HIV-1 serostatus were examined using χ 2 test or Fisher's exact test as appropriate. Second, to identify the independent predictors of HSV-2 discordance compared to negative concordance and the independent predictors of HSV-2-positive concordance compared to discordance, we performed two separate backward stepwise logistic regression analyses, with an exit significance level of 0·15. Sociodemographic characteristics and sexual behaviours were considered as potential predictors. The interaction term of frequency of sex in the past year and length of sexual relationship was also assessed in the model for HSV-2-positive concordance compared to HSV-2 discordance. Here, we combined the individual demographic and sexual behaviuoral variables of male and female spouses of a couple into one couple-level variable to avoid correlation. Variables with a P value <0·10 in the univariate models were selected for entry into stepwise logistic regression models. If one category of a dummy variable was included in the final model, the other category was forced into the final model.
As of 31 June 2010, a total of 1618 HIV-1-discordant couples were eligible. Of these, 1167 HIV-1-infected spouses and 1052 HIV-1-uninfected spouses completed the baseline survey and HSV-2 testing, including both spouses of 954 couples, 213 HIV-1-infected spouses and 98 HIV-1-uninfected spouses alone. For the present analyses, only the 954 couples with both spouses participating in the baseline survey and receiving HSV-2 testing were included.
Of the 954 participating couples, 78·3% had an HIV-1-infected male spouse and 21·7% had an HIV-1-infected female spouse. For the 954 male spouses, 19·6% were aged 16–29 years, 64·9% were of non-Han ethnicity, 64·7% were illiterate or had only primary school education and 86·6% were in their first marriage. For the 954 female spouses, 32·1% were aged 14–29 years, 72·4% were of non-Han ethnicity, 70·5% were illiterate or had only primary school education, and 82·1% were in their first marriage.
As shown in Table 1, 20·7% had been in the sexual relationship with the study partner for <3 years, whereas 79·3% had been in the relationship for ⩾3 years. In terms of sexual behaviours, slightly more than a half of the couples reported having had sex with the study partner ⩾4 times per month in the year prior to enrolment. Less than a quarter (22·9%) of the couples reported having had any unprotected sex with the study partner in the year prior to enrolment. Lifetime commercial sex, lifetime casual or anonymous sex, and syphilis infection, were all more common in HIV-1-infected male and female spouses than in HIV-1-uninfected male and female spouses (P < 0·05).
HSV-2 prevalence, discordance and concordance
HSV-2 prevalence was 35·2% in HIV-1-infected spouses and 27·7% in HIV-1-uninfected spouses, resulting in an overall prevalence of 31·4%. The highest HSV-2 prevalence (58·5%) was observed in HIV-1-infected females, compared to the lower prevalence in HIV-1-infected males (28·8%), HIV-1-uninfected females (28·3%), and HIV-1-uninfected males (25·6%) (P < 0·05) (Table 1).
Regarding HSV-2 status at the couple level, 405 (42·4%) couples were affected by HSV-2. These included 195 (20·4%) HSV-2-concordant-positive couples, 73 (7·6%) HSV-2-discordant couples where the male spouse was HSV-2 positive, and 137 (14·4%) HSV-2-discordant couples where the female spouse was HSV-2 positive.
Correlates of HSV-2 discordance compared to HSV-2-negative concordance
Table 2 shows results of univariate and multivariate analyses for variables associated with HSV-2 discordance compared to HSV-2-negative concordance. In the final multivariate model, compared to HSV-2-negative concordance, HSV-2 discordance with an HSV-2-positive male spouse was significantly associated with characteristics of the male spouse including Han ethnicity and being in a second marriage, but not with any sexual behavioural variables, syphilis or gender of the HIV-1-infected spouse.
cOR, Crude odds ratio; aOR, adjusted odds ratio.
* Only male spouses were included in the regression analysis.
† Only female spouses were included in the regression analysis.
In the final multivariate model, compared to HSV-2-negative concordance, HSV-2 discordance with an HSV-2-positive female spouse was significantly associated with characteristics of the female spouse including Han ethnicity, having engaged in commercial sex, having a sexual relationship of <3 years and being HIV-1 infected.
Correlates of HSV-2-positive concordance compared to HSV-2 discordance
Table 3 shows results of univariate and multivariate analyses for variables associated with HSV-2-positive concordance compared to HSV-2 discordance. In the final multivariate model, compared to HSV-2 discordance, HSV-2-positive concordance was significantly and positively associated with an education level of middle school or higher for both spouses, a sexual relationship of ⩾3 years, more frequent sex with study partner in the past year (⩾4 times/month) and having an HIV-1-infected male spouse. The interaction between frequency of sex and length of sexual relationship (⩾3 years) was marginally statistically significant (P = 0·056), with adjusted odds ratios (aORs) decreasing as the years in the sexual relationship increased.
cOR, Crude odds ratio; aOR, adjusted odds ratio.
This is the first study with a large sample size to examine the prevalence and risk factors of HSV-2 infection in HIV-1-discordant couples in China. We found that 42·5% of HIV-1-discordant couples were affected by HSV-2; lower than 75·3% of couples found in a previous study conducted in Kenya [Reference Muiru14]. This discrepancy may be partially due to the different predominant transmission routes of HIV in the two study areas. As suggested by surveillance reports and previous studies, drug injection is the predominant mode of HIV-1 transmission in the study area [Reference Jia18], whereas heterosexual transmission remains the most prominent mode of HIV transmission in Kenya [Reference Gouws19]. In addition, the overall prevalence of HSV-2 (31·4%) in our study is lower than prevalences reported in female commercial sex workers (33·2–68%) in other areas in Yunnan province [Reference Wang20, Reference Ngo21], but higher than in rural residents (13·4%) in Taizhou, Zhejiang province [Reference Lin22] and male general migrants (5·5%) in Shanghai [Reference He23]. These results, combined with the synergistic effect between HSV-2 and HIV-1 infections, demonstrate the importance of screening and treatment of HSV-2 infection in HIV-1-discordant couples.
Consistent with previous reports [Reference Muiru14], we observed that female spouses had a higher prevalence of HSV-2 than males, even though males had a higher HIV-1 prevalence than females. Women are at greater risk of HSV-2 infection because of higher efficiency of HSV-2 transmission from men to women compared to transmission from women to men [Reference Gottlieb24]. This suggests that HSV-2 prevention programmes should target women in particular. With regards to ethnicity, our results indicate that Han ethnic participants were more likely to be infected with HSV-2 than non-Han ethnic minorities regardless of their gender. Our previous study indicated that compared to Han ethnicity, these minorities had a higher percentage of HIV infection through injection drug use but had a lower percentage of HIV infection through sexual contact [Reference He25]. HSV-2 is primarily transmitted through sexual activity [Reference Muiru14], and there is a repeatedly observed strong association between HSV-2 infection and sexual risk behaviours [Reference Patnaik26]. This reinforces the need of sexual risk reduction interventions in the Han ethnic population. Furthermore, the association of second marriage with HSV-2 infection is not surprising because a higher number of sex partners increases the risk of HSV-2 infection [Reference Patnaik26, Reference Weiss27].
We observed that lifetime commercial sex by the female spouse was significantly associated with HSV-2 discordance where the female spouse was HSV-2 positive, compared to HSV-2-negative concordance. This suggests that commercial sex (i.e. selling sex) was one of the main routes for females to become infected with HSV-2 from a partner outside the marital partnership, as our data indicated that 9·2% of HIV-1-infected female spouses have ever had commercial sex. This is consistent with the route of sexual transmission of HIV-1 in females in the study area [Reference Jia18], highlighting the urgent need for HSV-2 and HIV-1 control and prevention in female sex workers in this area. Nevertheless, we did not find this association in male spouses. It may be attributed to the relatively small sample size of HSV-2-discordant couples where the male spouse was HSV-2 positive in our study, or/and the underreporting of outside sexual engagements by male spouses.
It is not surprising that HSV-2-discordant couples with an HSV-2-positive female spouse were more likely to have an HIV-1-infected female spouse than HSV-2-concordant negative couples. This is consistent with our data that HIV-1-infected female spouses had a higher prevalence of HSV-2 than HIV-1-uninfected female spouses. Higher prevalence of HSV-2 was also observed in HIV-1-infected male spouses than in HIV-1-uninfected male spouses. This indicates a tendency of HIV-1 and HSV-2 co-infection in HIV-1-discordant couples, particularly in female spouses. A similar pattern has also been observed in a previous study [Reference Muiru14]. HSV-2-concordant positive couples were more likely to have an HIV-1-infected male spouse compared to HSV-2-discordant couples. The possible explanation was that there were two types of HSV-2-discordant couples, and HSV-2-discordant couples with an HSV-2-positive female spouse had the lowest percentage of HIV-1-infected male spouses (i.e. highest percentage of HIV-1-infected female spouses) compared to other types of couples (data not shown) .
Consistent with a previous study [Reference Wald28], we observed that frequency of sex was associated with an increased likelihood of HSV-2-positive concordance. It is important to note that the interaction between the length of sexual relationship and frequency of sex was statistically significant, with aORs decreasing as the length of the sexual relationship increased. The frequency of sex in the previous year was not as strongly associated with HSV-2 in couples with a sexual relationship of ⩾3 years as those with a sexual relationship of <3 years. This suggests that HSV-2 transmission within the couple is more likely to occur in the first year of the sexual relationship. Therefore, prevention efforts should begin as early as possible in the relationship in order to reduce the likelihood of HSV-2 transmission within couples.
Several limitations of this study should be noted. First, all participating couples were recruited from a rural area in China; our findings may not be generalizable to other areas of China. Second, because this was a cross-sectional study design, the causal relationship between associated factors and outcomes could not be confirmed. Third, like most studies of sexual behaviours, this study was also potentially subject to social desirability bias, and frequency of unprotected sex and outside sexual partners may be underestimated. However, this was not a problem for judging the magnitude of HSV-2 infection in the study population. Fourth, HSV-2-positive concordance may reflect acquisition from outside the partnership or HSV-2 transmission within the couple, which limits the ability to identify the factors that affect HSV-2 transmission within the couple. In addition, because multiple testing bias remains a concern in multivariate models when interpreting individual covariates, our results should be interpreted with caution.
In conclusion, our observation of high HSV-2 prevalence in HIV-1-discordant couples highlights the need for HSV-2 prevention and treatment efforts in this population in order to reduce HSV-2 acquisition and transmission. Specifically, routine HSV-2 screening, risk reduction counselling and suppressive therapy for HSV-2 infection should be given to this population, particularly females. The implementation of HIV-1 pre-exposure prophylaxis in HIV-1-discordant heterosexual couples is also recommended, which has been recently shown to significantly reduce the risk for HSV-2 acquisition [Reference Celum29], besides offering protection against HIV acquisition. Our findings also emphasize the importance of implementing prevention interventions early in couples' relationships.
This study was supported by the National Science and Technology Major Project on Prevention and Treatment of Major Infectious Diseases including AIDS and Viral Hepatitis from the Chinese Ministry of Health (grant no. 2008ZX10001-016), and the National Natural Science Foundation of China (grant no. 81072345).
DECLARATION OF INTEREST