Methods

The study was conducted at the Sydney Sexual Health Centre, a free public inner city clinic in Sydney, Australia. All cases of first-episode anogenital herpes were identified from the clinic electronic database for the period 1992–2006. Cases were defined as patients with anogenital lesions with microbiologically confirmed HSV and no previous diagnosis of anogenital herpes. Cell culture was used until August 2004, after which a validated in-house PCR assay was used. Proforma medical record review was used to confirm the type and anatomical site of HSV infection for each case. The following variables were extracted from the electronic database: age, sex, HIV status, sex of sexual partners in the past 12 months and reported anal or vaginal sex in the previous 3 months including condom use. As only eight men (all MSM) were found to be have HIV, this variable was excluded from further analysis.

MSM were defined as men who reported sex with men only or with women and men in the previous 12 months. Women who have sex with women (WSW) were defined as women who reported sex with women only or with women and men in the previous 12 months. Heterosexual men and women were those reporting sex with opposite sex partners only in the previous 12 months.

Statistical analysis was performed using Stata version 10.0. Trend in the proportion of first episode anogenital herpes due to HSV-1 was determined using the Mantel–Haenszel χ² test for trend. Stratified analysis of trends was performed by age greater or less than median and gender of sexual partners. Associations with the overall proportion of HSV-1 were determined for age group, gender of sexual partners, anatomical site, reported anal or vaginal sex and use of condoms by calculating odds ratios (OR) with 95% CI. A logistic regression model was created to determine variables predicting the proportion of HSV-1 controlling for factors found to have a p value less than 0.1 on univariate analysis. The model was developed using forward stepwise logistic regression. WSW were excluded from the stratified analysis and logistic regression model due to the small sample size. Ethics approval was granted by the South Eastern Sydney Illawarra Area Health Service Human Research Ethics Committee.

Results

Between 1992 and 2006 there were 4440 clinically defined first episodes of anogenital herpes, of which 1845 (42%) were microbiologically confirmed HSV and 653 (35%) of these were HSV-1 (table 1). There were three cases with both HSV-1 and HSV-2 isolated concurrently; for the purposes of analysis these were treated as separate cases.

Trends in the proportion of first-episode anogenital herpes due to HSV-1

The overall proportion due to HSV-1 increased from 29% in 1992–4 to 42% in 2004–6 (OR per 3-year band 1.19; 95% CI 1.11 to 1.27; fig 1). Although the increase in the HSV-1 proportion remained when stratified by sex and gender of sexual partners, this only reached statistical significance for heterosexual women (p trend <0.01; table 1 and fig 1). Too few WSW were diagnosed (n  =  41), thus they were excluded from stratified trend analyses. A stratified analysis by age and gender of sexual partners found that the proportion due to HSV-1 only significantly increased in the case of MSM and heterosexual women less than the median age (28 years; fig 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Proportion of first-episode microbiologically confirmed anogenital herpes due to herpes simples virus type 1 (HSV-1), 1992–2006.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Proportion of first-episode microbiologically confirmed anogenital herpes due to herpes simplex virus type 1 (HSV-1) in those under 28 years, 1992–2006.

Predictors of first-episode anogenital herpes due to HSV-1

Overall, MSM were more likely to have first-episode anogenital herpes due to HSV-1 than heterosexual men (52% vs 27%, OR 2.89; 95% CI 2.15 to 3.87) or women (37%, OR 1.84; 95% CI 1.39 to 2.45) and equally likely as WSW (48%, OR 1.14; 95% CI 0.59 to 2.20). The proportion of cases who were MSM increased from 6.5% to 21.4% over the period (p trend <0.001). First-episode anogenital herpes was more likely to be caused by HSV-1 in younger compared with older people of all risk groups (table 1).

The proportion due to HSV-1 was higher for first-episode anal compared with genital herpes (48% vs 34%, OR 1.78; 95% CI 1.30 to 2.44; table 1). MSM were much more likely to have anal disease than heterosexual men (46% vs 2%, OR 62; 95% CI 32 to 121), heterosexual women (46% vs 7%, OR 15; 95% CI 10 to 23) and WSW (46% vs 4%, OR 7.8; 95% CI 2.70 to 22.48; data not shown). There was a non-significant increased likelihood of anogenital herpes due to HSV-1 in those not reporting vaginal or anal sex in the previous 3 months (OR 1.39; 95% CI 0.98 to 1.97) and in those reporting consistently using condoms for anal and vaginal sex in the previous 3 months (OR 1.16; 95% CI 0.94 to 1.43; table 1).

Multivariate logistic regression analysis

Over time, the increasing proportion of first-episode anogenital herpes due to HSV-1 remained significant after adjustment for age, gender of sexual partners and anatomical site (OR per 3-year band 1.16; 95% CI 1.08 to 1.25) (table 2). When the analyses were stratified by age, however, this increasing trend was only significant in younger MSM (OR 1.58; 95% CI 1.14 to 2.19) and younger heterosexual women (OR 1.30; 95% CI 1.14 to 1.48). No significant increasing trend was found in patients over 28 years.

Discussion

In this study we have demonstrated that the proportion of first-episode anogenital herpes due to HSV-1 increased significantly between 1992 and 2006, particularly in younger MSM and among heterosexual women. The proportion of first-episode anogenital herpes due to HSV-1 was significantly higher in MSM than in both heterosexual women and other men. HSV-1 was relatively more likely to be the cause of anal compared with genital disease and anal disease was far more common in MSM than other groups.

Two mechanisms have been proposed to explain the shift towards HSV-1: (1) a reduction in the childhood acquisition of HSV-1 leading to more susceptible young adults and (2) an overall or relative increase in the incidence of oral sex. There are data to support a decrease in the childhood acquisition of HSV-1; a US population-based study demonstrated a 14% decrease in HSV-1 seroprevalence in those younger than 20 years between 1988–94 and 1999–2004,9 and in the UK the seroprevalence of HSV-1 in 10–14-year olds declined from 34% in 1986–7 to 24% in 1994–5.10 Only one population-based HSV seroprevalence study has been conducted in Australia, indicating that 76% of adults over 25 years of age were HSV-1 positive.11 The prevalence of HSV-1 was similar in the Health in Men cohort but, notably, only 54% of MSM under 25 years of age were HSV-1 positive.8

We could find no significant association between HSV type and recent sexual behaviour other than the gender of sexual partners, but our database contains no consistent data on oral sexual practices. Others have demonstrated that recent oral sex is associated with symptomatic genital HSV-1 in heterosexual individuals.6 Among MSM in Sydney both oropenile and oro-anal practices have been associated with incident HSV-1 infection.8 Furthermore, oro-anal sex between MSM in Sydney appears to have increased during our study period. In repeated cross-sectional studies in Sydney in 1986–7, 1993–5 and 2001–3, the proportion of MSM reporting receptive oro-anal sex in the previous 6 months was 24%, 60% and 69%, respectively.12 This large increase in the proportion of MSM reporting oro-anal sex may at least partly explain the shift towards HSV-1 in first-episode anal lesions.

Fewer than half of the clinically defined cases were confirmed microbiologically. The use of culture for the majority of the study period would have increased the proportion of unconfirmed cases. As there are no data to suggest that cell culture and PCR differ in relative sensitivity for anogenital HSV-1 or HSV-2 this should not introduce any bias. As we used a clinical definition of first-episode herpes without serological evidence, some cases of recurrent herpes are likely to have been misclassified as first episodes. As HSV-2 causes more frequent recurrences,13 this would probably underestimate the importance of HSV-1. Conversely, as anogenital HSV-1 infections are more likely to be true primary infections (no previous HSV infection) they are likely to be more clinically severe and drive patients to seek healthcare. This would result in a bias towards HSV-1 infection, but seems unlikely to affect time trends.

This increasing propensity for HSV-1 to cause first-episode anogenital herpes has both clinical, public health and research implications. For some patient populations, for example young MSM and young women, HSV-1 may account for the majority of cases. For these patients the type-specific diagnosis is important for counselling about prognosis and about modes of transmission. In terms of public health, it is clear that interventions to prevent anogenital herpes, including vaccines, need to consider HSV-1 as well as HSV-2. For researchers, positive HSV-2 serology is often used as a surrogate for anogenital herpes.14 At least in our clinical population, this concept increasingly needs to be re-evaluated.