Background:  This multicenter prospective cohort study (October 1994 to September 2002) determined incidence and predictors of genital warts and vulvar intraepithelial neoplasia (VIN) and defined rates of progression and regression of grade 1 cervical intraepithelial neoplasia (CIN1) among women with HIV. Study.

Methods:  Every 6 months, women with HIV underwent genital examination and Pap smear, with colposcopy and biopsy by standard protocol, along with CD4 and HIV RNA measurement. Human papillomavirus (HPV) typing was performed at baseline from cervicovaginal lavage. HAART was defined according to criteria of the DHHS/Kaiser Panel. Data from 1562 HIV seropositive and 469 seronegative women without warts or VIN at baseline were included in analysis of vulvar disease, and CIN 1 was identified in 202 HIV seropositive and 21 HIV seronegative women during up to 6.7 years of observation. For CIN1 analysis, women were censored at the time of cervical therapy. Statistical analysis was by log-rank test, Kaplan Meier curves, and Cox multivariate analysis.

Results:  The incidence of genital warts among HIV seronegative women was 1.31/100 person years (95%CI:  0.92, 1.83), versus 5.01 cases/100 person-years (95%CI:  4.52, 5.53) among seropositive women (p <0.001). The incidence of VIN among seronegative women was 1.31/100 person-years (95%CI:  0.91, 1.82), versus 4.67/100 person-years (95%CI:  4.20, 5.18) among seropositives (p <0.001). In multivariable analysis, incident warts in HIV-infected women were associated with HAART (RH = 0.74, 95%CI:  0.64, 0.86), CD4 count (RH = 0.91/100 cell/mm³ increase, 95%CI:  0.86, 0.96), clinical diagnosis of AIDS (RH = 1.25, 95%CI:  1.00, 1.25), abnormal Pap smear (RH = 2.18, 95%CI:  1.73, 2.75), high or medium risk HPV types (RH = 1.91, 95%CI:  1.48, 2.47) and low risk types (RH = 1.48, 95%CI:  1.10, 2.00) compared to no HPV, smoking (RH = 1.43, 95%CI:  1.09, 1.88), having 1 child (RH = 1.54, 95%CI:  1.11, 2.13) compared with no children, and age (RH = 0.74/10 years, 95%CI:  0.64, 0.86). Incident VIN was linked to abnormal Pap smear (RH = 16.03, 95%CI:  11.33, 22.69) and high or medium risk HPV types (RH = 1.37, 95%CI:  1.06, 1.77) compared with no HPV and was linked inversely to HAART RH = 0.65, 95%CI:  0.49, 0.88, CD4 count (RH = 0.92, 95%CI:  0.87, 0.97), and age (RH = 0.85/10 years, 95% 95%CI:  0.73, 0.99). In a separate multivariable model including HIV seronegative women, HIV seropositivity increased the risk of warts (RH = 2.62, 95%CI:  1.78, 3.86) and VIN (RH = 1.52, 95%CI:  1.02, 2.26). Progression of CIN1 was observed in only 8 HIV⁺ and no HIV^- women. Regression occurred in 66 HIV⁺ women over 476.6 person-years of HIV⁺ (14%), versus 14 HIV^- women over 40.2 person-years (35%, p = 0.0003). After adjusting for age, smoking, HPV risk group, and ethnicity, regression was less likely among women with HIV (RH = 0.31, 95%CI:  0.16, 0.57, p < 0.001). Among HIV-infected women, regression was less likely among women with high risk HPV (RH = 0.51, 95%CI:  0.26, 0.99) but not low risk HPV (H.R. 0.93, 95%CI:  0.44, 1.96) compared with HPV-uninfected women. Regression was least likely in Latinas (RH = 0.41, 95%CI:  0.18, 0.94, p = 0.03) and marginally less likely in African Americans (RH = 0.56, 95%CI:  0.30, 1.04, p = 0.07) compared with whites. Age, smoking status, HIV viral load, CD4 lymphocyte count, and HAART were not associated with regression.

Conclusions:  Genital warts, VIN, and CIN1 are common among women with HIV, while progression of CIN1 is uncommon. HAART decreases incidence of both warts and VIN, while HPV risk type rather than antiretroviral therapy determines likelihood of regression of CIN1.

Keywords: genital warts; vulvar dysplasia; HAART