Dukers et al. Respond to "Sexual Practices Associated with HHV8 Infection"

Abstract

Although we pursued comprehensiveness in listing potential biases in our paper (1), Martin and Osmond (2) correctly noted an additional form of bias caused by the difficulty to accurately pinpoint the moment of infection. Although this may be a problem, recent observations in our cohort suggest a relatively short window-period. Among seroconverters with transient human herpesvirus 8 (HHV8) viremia before or at seroconversion, the period between detection of virus and detection of antibodies was short and was only rarely longer than 3 months (J. Goudsmit, University of Amsterdam, unpublished manuscript). Therefore, the extent to which sexual practices are measured after the actual moment of infection will probably be limited. Some bias cannot be excluded, but it is not likely that this leads to important spurious associations. Although inaccurate reporting of behavior might still influence our results (as is the case in any study design), our study is more sensitive in estimating an association with specific sexual behavior compared with previous studies, as we included a large number of incident cases and as most other studies were cross-sectional. Although we identified orogenital sex as the most important risk factor for HHV8 seroconversion, which was previously unrecognized, transmission by other routes cannot be ruled out. It was suggested that controlling for human immunodeficiency virus (HTV) might attenuate the potential risk for anogenital receptive sex. However, whether or not HTV was included in the multivariate model, the estimated risk for anogenital receptive sex (corrected for all other sexual techniques) was far from statistical significance. Still, assuming that transmission by anogenital and/or oroanal sex is possible, statistical significance may not be reached when transmission efficacy of these techniques is very low, as suggested by Martin and Osmond. Although we agree with their statement that because of such universal practices as kissing and penile-oral intercourse saliva may transmit inefficiently, the transmission efficacy of receptive anogenital or insertive oroanal sex may even be lower as less or no virus is detected in semen and feces, respectively (3). Since the start of the HTV/acquired immunodeficiency syndrome (AIDS) epidemic, prevention campaigns for homosexual men have focused on anogenital sex. As a result, homosexual men reduced their number of partners and the frequency of unprotected anogenital contacts, and the HTV incidence declined strongly. However, these prevention campaigns appeared to have had less impact on sexual practices other than anogenital receptive sex as, for example, in our cohort the frequency of unprotected orogenital contact has remained constantly high over the years. Sexual practices other than anogenital receptive sex can be of significant importance in the transmission of sexually transmitted diseases, as is well known for hepatitis A and B, for example, and as we have now suggested for HHV8. Therefore, we fully agree with Martin and Osmond that the prevention message for homosexual men should not be targeted toward the risk of a specific sexual technique but toward safe sex in general, including safe oral sex.