Determinants of human papillomavirus (HPV)-16 serological conversion and persistence were assessed in a population-based cohort of 10 049 women in Guanacaste, Costa Rica. Serologic responses to HPV-16 were measured in 7986 women by VLP-based enzymelinked immunosorbent assay at both study enrolment (1993/94) and at 5 -7 years of follow-up. Seropositive women were defined as X5 standard deviations above the mean optical density obtained for studied virgins at enrolment (n ¼ 573). Seroconnversion (n ¼ 409), persistence (n ¼ 675), and clearance (n ¼ 541) were defined based on enrolment and follow-up serology measurements. Age-specific distributions revealed that HPV-16 seroconversion was highest among 18-to 24-year-old women, steadily declining with age; HPV-16 seropersistence was lowest in women 65 þ years. In age-adjusted multivariate logistic regression models, a 10-fold risk increase for HPV-16 seroconversion was associated with HPV-16 DNA detection at enrolment and follow-up; two-fold risk of seroconversion to HPV-16 was associated with increased numbers of lifetime and recent sexual partners and smoking status. Determinants of HPV-16 seropersistence included a 1.5-fold risk increase associated with having one sexual partner during follow-up, former oral contraceptive use, and a 3-fold risk increase associated with HPV-16 DNA detection at both enrolment and follow-up. Higher HPV-16 viral load at enrolment was associated with seroconversion, and higher antibody titres at enrolment were associated with seropersistence. (Ho et al, 1995). Moreover, not all women infected with HPV seroconvert; only about half of all HPV DNA-positive women test positive for corresponding type-specific antibodies using available assays (Kirnbauer et al, 1994;Le Cann et al, 1995). For those women who have seroconverted, however, detection of serum antibodies to HPV capsids is a valid marker of current and past type-specific HPV exposure (Wideroff et al, 1995(Wideroff et al, , 1999Carter et al, 1996;Sasagawa et al, 1998;Touze et al, 2001).We previously reported population-based seroprevalence of HPV-16, -18, -31, and -45 in our 10 000 women population-based study in Costa Rica at enrolment; we confirmed the waning detection of HPV antibodies with age and the determinants of seroprevalence to include increasing lifetime number of sexual partners and smoking (Wang et al, 2003), as similarly reported in other prevalence studies (Stone et al, 2002;Nonnenmacher et al, 2003).To extend our cross-sectional findings, we have now completed HPV-16 serology measurements at a second time point, after 5 -7 years of follow-up. Human papillomavirus 16 was selected because it accounts for the majority of cervical cancers worldwide (Munoz et al, 2003) and is the focus of immunology and vaccinology research (Lowy and Frazer, 2003). To further our understanding of HPV serology in the natural history of cervical cancer, we report here the epidemiologic characteristics and determinants of HPV-16 serologic conversion and persistence in our Guanacaste, Costa ...