Factors predicting the persistence of genital human papillomavirus infections and PAP smear abnormality in HIV-positive and HIV-negative women during prospective follow-up.Int J STD AIDS. 2003 Jun; 14(6):417-25.IJ
As part of an extensive multi-institutional DIANAIDS study focused on assessing the risk factors, natural history, diagnosis and follow-up of genital human papillomavirus (HPV) infections in HIV-infected women, the present communication reports a sub-cohort of 142 women (89 HIV+ and 48 HIV-), followed-up for a mean of 14.07 (+/-10.84) months to analyse the factors predicting the persistence and clearance of HPV infections (polymerase chain reaction [PCR] and sequencing) and cervical Papanicolaou (PAP) smear abnormalities, using both univariate (Kaplan-Meier) and multivariate (Cox) survival analysis. The appearance of new HPV infections during the follow-up was significantly more frequent in HIV-positive than in HIV-negative women, odds ratio (OR) 8.800 (95% confidence interval [CI]: 1.199-64.611), and also the clearance rate was significantly less frequent in HIV-positive than in HIV-negative women, 69.2% vs 22.8%, respectively (OR 0.330; 95% CI: 0.163-0.670). These two groups were also markedly different with respect to the clinical course of the cervical lesions, in the frequency of progressive disease (determined by PAP smear) was higher in HIV-positive group (12/89) than in HIV-negative women (2/52) (OR 3.506; 95% CI 0.816-15.055) (P = 0.055), in whom the disease regressed more frequently than in HIV-positive women (13.5% vs 7.9%) (OR 0.584; 95% CI 0.217-1.573). Using (1) HPV-positivity, (2) oncogenic HPV-type and (3) significant PAP smear abnormality at the end of follow-up as outcome measures, (1) was significantly (P < 0.001) predicted by the following variables in univariate analysis: age, mode of contraception, CD4 count, and HIV-positivity. The significant predictors of (2) were age and mode of contraception. The outcome measure (3) was significantly predicted by CD4 count, PAP smear abnormality and PCR status at entry. In the multivariate analysis, the significant independent predictive factors for HPV-positivity proved to be only the HIV status (P < 0.001), and PCR status at entry, p53 polymorphism at aa-72, oncogenic HPV type and significant PAP smear at entry remained independent predictors, with the significance level of P < 0.05. None of the significant predictors of oncogenic HPV type in univariate analysis retained their independent value in multivariate analysis. Oncogenic HPV type at entry proved to be an independent predictor of significant PAP smear (P < 0.05). The present results indicate that HIV-infected women, even on highly active antiretroviral therapy, demonstrate a more aggressive clinical course of cervical HPV infections, and fail to eradicate the disease more frequently than HIV-negative women. This persistence of HPV-positivity, oncogenic HPV type and significant PAP smear abnormality can be predicted by the results of PAP test and HPV typing in univariate analyses, and partly retain their independent predictive value also in multivariate analysis. Clearly, in addition to regular monitoring by PAP smear, HPV testing for the oncogenic HPV types seems to provide additional prognostic information in the management of cervical lesions in HIV-infected women.