HSV-2 seroincidence and the unintended pregnancy rate in young women receiving family planning services in San Francisco clinics over a 6-month period were high at 7.8 and 12.7 per 100 person-years, respectively. The rate of HSV-2 acquisition in women who became pregnant during the study period was 10.1 cases per 100 person-years. Seroincidence of HSV-2 has recently been reported as 7.3 cases per 100 person-years in a study of 100 sexually active females aged 14 to 17 years [12
]. A larger study found a HSV-2 seroincidence of 13.2 cases per 100 person-years in males and females aged 14 to 24 at inner city STD clinics in the mid 1990's [13
]. Our own data, along with these, are consistent with the increase in HSV-2 seroprevalence observed nationally as women move from adolescence into their late twenties [7
]. We are unaware of other studies demonstrating high rates of unintended pregnancy in a population of young women that also has high seroincident HSV-2.
Recent NHANES data showed higher HSV-2 seroprevalence associated with increasing age, younger age at first intercourse, number of lifetime sexual partners, and poverty [7
]. African American women also had relatively high HSV-2 seropositivity [7
]. While bivariate analysis of our data suggested several baseline behavioural and demographic predictors of HSV-2 seroincidence, multivariable analysis found African American race and multiple sex partners within the last six months increased the likelihood of new infection. The risks were high: African American women were nearly 2.5 times more likely to acquire HSV-2 infection than non-African American women and those with multiple partners were 75% more likely. Clinicians should be aware of the significant increased risk associated with these factors.
Our findings showed condom use at the last intercourse prior to enrolment was highly protective, associated with a 44% reduction in likelihood for acquiring HSV-2 after adjustment for confounders. This is consistent with a recent study by Wald et al showing that STD clinic patients reduced their risk for HSV-2 by 26% as they increased condom use frequency [14
]. Our study is at least the fourth to demonstrate that condom use was associated with decreased HSV acquisition and should encourage medical providers and policy makers to feel confident in recommending condom use to decrease HSV-2 transmission [13
We observed a high rate of unintended pregnancies in an urban, racially diverse group of young women. This concurs with findings from the National Survey of Family Growth for 2001 that showed poor, uneducated, minority women are at increased risk for unintended pregnancy [9
]. Interestingly, nearly half of these unintended pregnancies occurred in women reporting concurrent contraceptive use [9
]. While participants did not intend to become pregnant in the next six months at enrolment, they could have changed their minds after enrolment and intentionally conceived before the 6-month follow up appointment. Information regarding change in intention in the interim was not collected.
The role of HSV in the development of severe neonatal complications following maternal infection is well known [3
]. Brown et al recently reported a neonatal herpes infection rate of 1 in 3200 live births (31.25 per 100,000 live births) for several hospitals around Seattle, Washington [6
]. Others have reported incidences of 11 to 33 infections per 100,000 live births [16
]. Rates of neonatal herpes are highest in neonates born to recently HSV-infected women who were seronegative pre-pregnancy [6
]. Our data suggest that young women in San Francisco run a high risk of acquiring HSV-2 infection shortly before and during their pregnancies.
These findings highlight the need for multiple interventions to minimize acquisition of HSV-2 infection in the young urban female population. Possible effective interventions include targeted screening and vaccination among women seeking family planning services as these women are at high risk for unintended pregnancy, despite contraceptive use. Targeted screening could include women who are African American or who report multiple sex partners within the last six months. We think it is also important for providers who counsel women on contraceptive services and STI prevention to play an expanded role in counselling women about prevention of HSV-2 infection given the potential sequelae in pregnancy. While vaccine development requires extensive bench to bedside translational research, the other interventions listed are well within the grasp of existing public health institutions and future studies should evaluate their effectiveness in reducing the burden of new HSV-2 infections in young women at risk for unintended pregnancy.
Also pertinent to this discussion, it has been suggested that HSV-2 antibody screening may place unnecessary psychological stress on individuals, as there is typically no intervention indicated for asymptomatic cases [19
]. Ample research by our group and others, however, indicates that HSV-2 testing does not cause inappropriate levels of psychological distress and thus prevention strategies that incorporate routine HSV-2 testing may be warranted [21
]. At the least, the evaluation of such strategies should be a public health priority.
With regard to limitations, the main study, of which this is a sub-study, was designed to examine emergency contraception on reproductive health outcomes in women using high compliance contraceptives not intending to become pregnant in the next six months. These criteria resulted in the exclusion of half of the women initially screened. Women using methods requiring a lower degree of user compliance, including injectable contraception and IUDs, would be expected to have a lower unintended pregnancy rate. Because women using low compliance methods were excluded, the unintended pregnancy rate calculated in our study may be an overestimate of the actual rate. However, these women tend to use condoms less frequently and, as a result, their exclusion may have led to an underestimation of HSV-2 seroincidence in our study. A cohort assembled with incident HSV-2 infection and pregnancy as the sole primary research endpoints might have avoided these potential biases.
Fourteen percent of women did not complete HSV-2 testing at 6-month follow-up. While the majority (64%) of these women were lost to follow-up, those remaining completed telephone interviews at follow-up for the main study but were not available for repeat fingerstick blood testing for HSV-2 or pregnancy testing. It is unknown how this impacted the measures of incidence or of associations. An analysis of attrition for the main study showed no significant differences in baseline demographic traits, history of pregnancy or STIs, or contraceptive methods between those who completed the study and those lost to follow-up.
In addition, we relied on the self-report of sexual risk behaviours, which women may recall with varying degrees of accuracy. This is a limitation of our study and, indeed, of other studies seeking to characterize behavioural factors affecting the acquisition of sexually transmitted infections. By contrast, the use of both urine testing, self-report and medical chart review in our assessment of new pregnancies is a strength of our study, although it is possible that some women may have under-reported new pregnancies during the study period that were terminated before urine testing at follow-up.
Finally, although the rapid point-of-care test we used to identify HSV-2 infection has a documented high sensitivity (96%) and specificity (98%), it is not as accurate as Western blot. A few women may have been falsely identified as HSV-2 positive while others may have been falsely identified as HSV-2 negative, especially if they had low HSV-2 titre levels. This would result in non-directional misclassification and would bias our findings towards not finding significant predictors of HSV-2 infection. Therefore our findings may have been stronger if we had used a more accurate test.