Article Text
Abstract
Objectives To examine herpes simplex virus type 2 (HSV-2) incidence over four periods to age 38 in a birth cohort, and to compare risks for men and women, taking into account sexual behaviour.
Methods At ages 21, 26, 32 and 38, participants in the Dunedin Multidisciplinary Health and Development Study were invited to provide serum for HSV-2 serology, and information on sexual behaviour. HSV-2 incidence rates were calculated for four age periods, and comparisons made by sex and period, taking into account number of sexual partners.
Results By age 38, 17.3% of men and 26.8% of women had ever been seropositive for HSV-2. Incidence peaked for women from age 21 to 26 (19.1 per 1000 person-years) and men from age 26 to 32 (14.1 per 1000 person-years); it fell markedly for both from age 32 to 38 (5.1 and 6.8 per 1000 person-years for men and women, respectively). Overall risk was significantly higher for women: adjusted incidence rate ratio 1.9 (95% CI 1.4 to 2.7); the sex difference was most marked from age 21 to 26 (3.4, 95% CI 1.9 to 6.3).
Conclusions Our findings are consistent with a greater biological susceptibility to HSV-2 among women, and with the increasing risk to the early/mid-20s for women and late 20s/early 30s for men, being driven by an increasing pool of prevalent infection. The reduced risk in the mid-30s is consistent with declining infectivity of long-term prevalent infections.
Statistics from Altmetric.com
Introduction
Herpes simplex virus type 2 (HSV-2) infection remains the commonest cause of genital herpes in spite of a recent increase in genital herpes simplex virus type 1 infection.1 ,2 To develop effective sexually transmitted infection (STI) control programmes, it is necessary to understand how epidemics evolve in populations.3 The drivers of HSV-2 infection—infectivity, rate of exposure between infected and uninfected people, and duration of infection—are difficult to measure directly. Nevertheless, incidence within a generation followed over time, linked with data on sexual behaviour, provides insight into these drivers.
As HSV-2 is frequently subclinical, incidence is ideally obtained through repeated serological assessment. However, previous incidence studies have generally been short and undertaken on selected groups; neither of the two more representative samples examined incidence past age 30 or accounted for sexual behaviour.4 ,5 We previously examined HSV-2 incidence to age 32 in the Dunedin Multidisciplinary Health and Development Study (DMHDS) through repeated serological assessment with associated information on sexual behaviour.6 Consistent with other studies, incidence was higher among women, and rose with age; we conjectured this was due to increasing prevalence of HSV-2 in partners.4 ,5
Here we report follow-up of the cohort to age 38, allowing delineation of HSV-2 incidence to an older age than previously studied. The aims are to examine HSV-2 incidence over four age periods to age 38, compare risks for men and women taking into account sexual behaviour, and consider what influences this variation.
Methods
Participants are members of the DMHDS cohort, born between April 1972 and March 1973, in Dunedin, New Zealand. Duration between assessments increased with age, and information on sexual behaviour was collected using a computerised questionnaire at ages 21, 26, 32 and 38. Participants provided serum for a number of purposes including serological assessment of HSV-2 antibodies under ethical approval from the Otago and Southern Regional Ethics Committees.
HSV-2 antibody testing of serum was undertaken using an indirect enzyme linked immunosorbent assay at the Centre for Virus Research, Westmead Millennium Institute, as previously reported.6
Numbers of opposite and same-sex partners were combined at each assessment; although at age 21 the only same-sex partners asked about were anal intercourse partners for men. Age at first coitus was based on the first experience of vaginal intercourse or same-sex contact.
Prevalence at age 38 was among those who provided sera at that age, and cumulative incidence for those who tested positive at any assessment. Incidence rates were calculated using the Poisson distribution for four age periods. At age 21, those who had experienced coitus and were tested for HSV-2 were included; at subsequent periods, those who tested negative for HSV-2 at the previous assessment and experienced coitus were included.
Incidence rate ratios (IRR) were calculated using multilevel mixed-effects Poisson regression with robust SEs and years since first coitus to approximate years at risk: by age period according to sex, by sex according to age period, and then by sex and age period to determine the overall effect by sex. IRRs were recalculated including number of sexual partners as a covariate.
STATA V.12.1/SE was used for all analyses.
Results
Of the cohort survivors, 75.9%, 76.1%, 86.7% and 88.7% completed questions on sexual behaviour and provided serum at ages 21, 26, 32 and 38, respectively.
The seroprevalence of HSV-2 antibodies at age 38 was 14.0% (63/451) for men, significantly lower than for women (23.7% (107/451), p<0.001). The cumulative incidence at age 38 was also lower for men (17.3%) than women (26.8%), and slightly higher than the prevalence, suggesting some sero-reversion.
The incidence rate differed by age period and sex (table 1). For men, it rose slightly from first coitus to age 21 to the period from age 21 to 26, then more steeply from age 26 to 32, before dropping back from age 32 to 38. For women, it rose more steeply from first coitus to 21 to its peak incidence from age 21 to 26, remained high from age 26 to 32, before dropping steeply from age 32 to 38. For both sexes in each period, the incidence increased with number of sexual partners, except for men from first coitus to 21.
Comparison of incidence rates, adjusting for number of sexual partners, showed clear differences by age period and sex. For men, compared with first coitus to 21, the adjusted IRR was not significantly higher from age 21 to 26 (1.5, 95% CI 0.7 to 3.2), but was from age 26 to 32 (2.7, 95% CI 1.3 to 5.4); from age 32 to 38 it dropped and was similar to the initial period (1.1, 95% CI 0.5 to 2.7). For women, compared with first coitus to 21, it was markedly higher from age 21 to 26 (3.7, 95% CI 2.0 to 6.9) and from age 26 to 32 (2.9, 95% CI 1.5 to 5.6); from age 32 to 38 it dropped and was not significantly higher than the initial period (1.5, 95% CI 0.7 to 3.4). Comparing men and women, overall across all periods, the adjusted IRR was significantly higher for women than men (1.9, 95% CI 1.4 to 2.7). The difference was most marked from age 21 to 26 (3.4, 95% CI 1.9 to 6.3), and not significantly otherwise (1.4, 95% CI 0.6 to 3.2 from first coitus to 21; 1.4, 95% CI 0.8 to 2.3 from age 26 to 32; 2.0, 95% CI 0.9 to 4.8 from age 32 to 38).
The proportion of new infections occurring among women with one/no partner increased steadily across the four periods (p-for-trend <0.001), a pattern not seen for men.
Discussion
In this population-based birth cohort, infection with HSV-2 was common, occurring in over a quarter of women and nearly one in six men by age 38. Overall prevalence and incidence were higher for women than men, driven mainly by the much higher incidence for women in the early/mid-20s when risk peaked. Risk peaked for men in the late 20s/early 30s, and incidence fell markedly for both sexes in the mid-30s. Incidence rose with number of sexual partners in nearly every age period; accounting for partner numbers accentuated risk differences by sex and age.
Limitations include the relatively small sample, and lack of sufficient data on condom use and behaviour of partners.
The seroprevalence at age 38 was similar to those aged 30–39 in the most recent US NHANES, and slightly higher than for those aged 35–44 in a large representative Australian sample obtained in 1999–2000.7 ,8 Increasing incidence with age to the mid-20s for women, and the early 30s for men, is consistent with two incidence studies among relatively representative samples.4 ,5 Decreasing incidence in the 30s has not been shown before.
These patterns of HSV-2 acquisition by age can be understood in terms of biological, behavioural and epidemiological factors. HSV-2 is shed in genital secretions, often as subclinical episodes, that gradually become less common over time.1 The increasing incidence in the 20s compared with younger ages is not a consequence of increasing numbers of partners. Hence the most likely reason is a growing prevalence among partners. If infectivity remained constant, increasing incidence with age would be expected, but incidence actually decreased in the mid-30s. The most plausible explanation is that, though prevalence continued to rise, the average infectivity of partners with HSV-2 decreased due to a higher proportion of infections of longer duration where there is decreased frequency of viral activation, and asymptomatic genital shedding.9 Other less likely explanations include less frequent intercourse, greater avoidance of sex during clinical recurrences, more condom use or possibly a greater incidence of HSV-1.
Incidence was significantly higher for women than men overall, being strongest in the early/mid-20s and not significantly so at other ages; this drove their much higher prevalence at age 38. While the different ages of peak incidence among women probably relate to patterns of sexual mixing with women tending to have older partners, their higher overall risk across all age periods supports them having greater biological susceptibility.
Risk increased with numbers of partners as for other STIs.10 Nevertheless, an increasing proportion of infections occurred among women without multiple partners in the older age periods, with nearly half of all infections from ages 32 to 38 in this group; this pattern was not seen for men. For women, this is most simply explained by the increasing proportion without multiple partners (68% from age 32 to 38 vs 22% before 21). Their single partner may have acquired a new infection or had a recurrence, though non-sexual infection is theoretically possible.8 However, men of that age without multiple partners accounted for less than 10% of new infections, despite the proportion without multiple partners also increasing with age. The reason for this sex difference is not clear.
These data provide evidence both that prevalence drives HSV-2 incidence up to the late 20s/early 30s, and that declining infectivity of long-term prevalent infections results in a reduced risk of infection from the mid-30s. Of relevance clinically, peak incidence of HSV-2 infection occurs at older ages than bacterial STIs,10 and later in men than in women. Moreover, the diagnosis of HSV-2 should be considered even among those who report one partner, of particular relevance for pregnant women because of the risk of neonatal transmission.
Key messages
-
This longitudinal study of a birth cohort has previously shown herpes simplex virus type 2 (HSV-2) incidence is higher in women than men, increasing with age into the 20s.
-
Follow-up showed HSV-2 incidence peaked for women in their early/mid-20s and men in their late 20s/early 30s; for both sexes risk declined steeply in the mid-30s, even after accounting for number of sexual partners.
-
The reduction in risk after age 32 is consistent with declining infectivity for long-term prevalent infections.
-
Cumulative incidence by age 38 remained higher for women (27%) than men (17%).
Footnotes
-
Handling editor Jackie A Cassell
-
Acknowledgements The authors thank the study members and their families for their long-term involvement in the DMHDS, the study director Professor Richie Poulton and staff involved in collection of the data and other aspects of the study, and Dr Jane Morgan for helpful comments.
-
Contributors ND, CP and TvR designed the study and wrote the manuscript. AR coordinated the collection of the data and undertook the analysis. JT and ALC were responsible for the testing of the samples. All authors critically reviewed and contributed to the manuscript, and approved the final version.
-
Funding The Health Research Council of New Zealand funded this study through grant number 09/086 that did not have any involvement in its implementation or analysis.
-
Competing interests None.
-
Patient consent Obtained.
-
Ethics approval Otago and Southern Regional Ethics Committee as relevant for each phase.
-
Provenance and peer review Not commissioned; externally peer reviewed.