| Title | Claps | Level | Year | L/Y |
|---|---|---|---|---|
|
The Relationship between Condom Use and Herpes Simplex Virus Acquisition
10 auth. A. Wald, A. Langenberg, Elizabeth M. Krantz, J. Douglas, H. Handsfield, R. Dicarlo, ...
Context We need other means to reduce the risk for transmitting genital herpes (herpes simplex virus type 2 [HSV-2]). Are condoms effective? Content In a trial of an ineffective HSV-2 vaccine, 1843 participants were divided into 3 groups according t…
Context We need other means to reduce the risk for transmitting genital herpes (herpes simplex virus type 2 [HSV-2]). Are condoms effective? Content In a trial of an ineffective HSV-2 vaccine, 1843 participants were divided into 3 groups according to the frequency of condom use (for 0% to 25%, 25% to 75%, and >75% to 100% of sexual acts). Frequent condom users had fewer HSV-2 infections. Compared with participants in the next lowest category, participants in a category had a 26% lower risk for HSV-2 infection. Limitation In this observational cohort study, many unmeasured factors could also contribute to altered rates of HSV-2 acquisition. Conclusion Condom use is associated with a lower rate of acquisition of HSV-2. The Editors Genital herpes is a common sexually transmitted infection that can be transmitted during episodes of recurrent lesions and during subclinical shedding (1). In the absence of an effective vaccine, condoms have been routinely recommended for prevention of transmission, and a recent study showed that daily antiviral therapy also decreases the risk for transmission of herpes simplex virus type 2 (HSV-2) in discordant couples (2, 3). In a previous study of monogamous HSV-2discordant couples who were enrolled in an ineffective candidate HSV-2 vaccine trial, we showed that condoms protect women from HSV-2 infection (4). However, very few cases of genital HSV-2 occurred among men who were sexual partners of women infected with HSV-2, precluding definitive conclusions about the effectiveness of condoms for prevention of transmission to men. We present data from a concurrent trial of the candidate vaccine among HSV-2seronegative persons attending sexually transmitted disease clinics (5). A total of 1862 participants were enrolled in this study; 85 cases of genital herpes were documented in men, and 33 cases were documented in women. We analyzed the effect of condom use on HSV acquisition in this prospectively followed cohort of men and women. Methods Study Sample Participants included in this analysis took part in a randomized, double-blind, placebo-controlled efficacy trial of a candidate subunit HSV-2 vaccine that was subsequently shown to be ineffective (5). The trial involved 22 centers located at sexually transmitted disease clinics and enrolled 1862 participants. Initial serologic testing was done at screening; participants who were seronegative for HIV and HSV-2 and reported 4 or more sexual partners in the past year or 1 or more sexually transmitted diseases in the past year were eligible to enroll. The effectiveness of condom use among the 528 discordant couples enrolled in a parallel vaccine study was reported previously (4, 5). Participants were enrolled and followed for 18 months, during which they were evaluated at 11 study visits. At enrollment, we collected demographic information and information about sexual history. At each study visit, we took blood samples and recorded the following information about sexual history, which described behavior since the last visit: frequency of sexual activities, defined as vaginal or anal intercourse; frequency of condom use; number of partners; number of new partners; and number of partners with a known history of genital herpes. The information regarding number of partners was gender-specific. In addition, participants were counseled routinely about safer sexual behavior and were offered condoms at each study visit. Genitourinary signs and symptoms were evaluated as needed at additional interim visits. Laboratory Methods The Western blot assay done at the University of Washington, Seattle, Washington, established HSV serologic status at study entry and was used to document seroconversion (6). Type-specific cultures using standard techniques were done at local study sites. Statistical Analysis Acquisition of HSV-2 was defined by seroconversion on the Western blot assay or by a positive culture for HSV-2. Time to HSV-2 acquisition was defined as the number of days from screening to the first positive culture for HSV-2 or as the midpoint between the last negative result of the HSV-2 antibody test and the first positive result of the HSV-2 antibody test. In this analysis of condom use and HSV acquisition, we included the time from screening to enrollment in the study, whereas in the vaccine trial participants were followed beginning at enrollment. Thus, our report includes 109 participants who were not included in the efficacy analysis of the original vaccine trial. Twenty of these participants seroconverted to HSV-2 during the screening period before enrollment, and 89 were lost to follow-up after enrollment. Participants who did not acquire HSV-2 were censored at the last blood draw taken during the study or at enrollment if they did not report any sexual activity thereafter. Participants who reported no sexual activity for the entire time from screening to study termination were excluded from the analysis because they were not at risk for HSV-2 infection. Participants with follow-up longer than 65 days beyond the 18 months specified in the protocol (3%) were censored at day 605. Participants who were seronegative for HSV type 1 (HSV-1) and HSV-2 at screening were included in the analysis of HSV-1. Time to HSV-1 acquisition was defined as the number of days from screening to the first positive culture for HSV-1 or the midpoint between the last negative result of the HSV-1 antibody test and the first positive result of the HSV-1 antibody test. Participants who did not acquire HSV-1 were censored at the last blood draw or at enrollment if they did not report any sexual activity thereafter. Participants who reported no sexual activity for the duration of the study were excluded from the analysis of HSV-1. KaplanMeier curves, log-rank tests, and univariate and multivariate Cox regression models were used to determine baseline risk factors associated with HSV-2 acquisition. To relate sexual behavior to HSV-2 acquisition during the study, we constructed time-dependent covariate Cox regression models. The analysis time was divided into four 150-day intervals, and information about sexual history collected at interim visits was used to calculate covariate summaries for each period. Because continuous variables did not satisfy the assumption of a linear effect in the log hazard, they were categorized. Our choice for the cut-points was motivated by maintaining equal numbers of participants in each category (for example, age was split at the median value, 27 years), by consistency with observed risk patterns, or by interpretation considerations. Frequency of sexual activity was expressed as the average number of sexual acts per week in the time period, calculated by averaging the reported estimates over the visits for each interval. This average was then categorized as greater than 2 versus 2 or fewer to correspond to observed risk patterns. Use of condoms during the study period was described categorically in each interval (used for 0% to 25%, for 25% to 75%, or for >75% of sexual acts). This grouped linear parameterization was chosen to remain consistent with published analyses (4) while allowing a doseresponse relationship, assuming constant change in the risk with increasing category of condom use. The use of condoms was not evaluated during intervals for which the participants did not report any sexual activity. The number of partners reported was summarized for each period and was modeled in a binary fashion. Partner cut-points were chosen for interpretation reasons to describe ways in which this patient group may differ from monogamous couples who were studied in a previously published report addressing condom use and infection with HSV (4). Total number of partners was modeled as more than 1 versus 1 or fewer, and both new partners and partners with a history of genital herpes were modeled as any versus none. These analyses did not adjust for receipt of placebo versus receipt of vaccine, because this factor was not statistically significant in acquisition of HSV and did not influence the covariates of interest for this study. An interaction term between condom use and gender was used to check the hypothesis of a difference in the effect of condoms by gender and to provide gender-specific estimates of condom use. Two-sided P values for model covariates were calculated by using the likelihood ratio test. The same methods were used to explore baseline risk factors and time-varying risk factors for time to infection with HSV-1. Poisson regression was used to provide P values for comparisons involving incidence rates. Tests for changes in sexual behavior with time used generalized estimating equations. Statistical analyses were done by using Stata statistical software (version 8.1, Stata Corp., College Station, Texas). Role of the Funding Source The funding for the analyses for this study was provided by federal grants; design, data analysis, and interpretation were done at the University of Washington. The initial clinical trial was funded by Chiron Corporation. This study was supported in part by National Institutes of Health Herpes Program Project Grant AI-30731 and Centers for Disease Control and Prevention Research Initiative UR6/CCU017828-02. Results Of the 1862 participants who enrolled for the vaccine trial, 19 did not report any sexual activity during the entire study and thus were excluded from this analysis. The remaining 1843 participants included 1365 men and 478 women. The median age of the participants was 27 years. Sixty-two percent were white, 32% were African American, and 6% were people of other races; 1184 participants (64%) were seropositive for HSV-1 at study entry. Most men and women qualified for the study by reporting 4 or more partners in the past year (66% of men, 70% of women); some reported 1 or more sexually transmitted diseases in the past year (12% of men, 19% of women); and the remainder met both criteria (22% of men, 1
Published in
Annals of Internal Medicine
|
7
|
7 | 2005 |
Social Media Posts