Factors predicting persistence of high-risk human papillomavirus (HPV) infections in women prospectively followed-up in three New Independent States (NIS) of the former Soviet Union.

Background: We completed an analysis of the factors predicting the persistence of high risk (HR) HPV infections in women participating in a multicenter screening trial in three NIS countries.

Methods: The 543 baseline HR HPV-positive women included in this analysis are derived from a sub-cohort of 887 women who were prospectively followed-up for a mean of 21.6 months (range: 0.5-42.9) as a part of a multi-center screening study in three NIS countries (the NIS cohort study; n = 3,187 women). Of these 543 women, 273 showed persistent HR-HPV in serial Hybrid Capture II (HCII) testing during the follow-up (Group 1), whereas 270 women cleared their infection (Group 2). These two groups were compared with their epidemiological, clinical, and virological data (HCII, PCR) to disclose the factors predicting persistent HR-HPV infection.

Results: Women with persistent HR-HPV infections were significantly younger (27.3 yrs) than those who cleared their infection (29.1 yrs) (p = 0.006), and their follow-up time was shorter; 14.1 and 21 months, respectively (p = 0.0001). Both variables were treated as confounders in the multivariate analyses. Of the 66 recorded epidemiological variables, only being a current smoker proved to be an independent predictor (OR 1.693; 95% CI 1.114-2.573; p=0.014). Baseline colposcopy, biopsy or Pap smear did not predict HPV persistence, whereas an incident or persistent abnormal Pap during the follow-up were independent predictors in a multivariate model (p = 0.005), together with the high viral load (HCII RLU/CO at 100 pg/ml cut-off), and HR HPV positive PCR test (p = 0.0001). When all significant variables were entered in the regression model, only the follow-up time (OR 0.950, 95% CI 0.924-0.976; p = 0.0001) and HR-HPV positive PCR (OR 4.169, 95% CI 1.741-9.987; p = 0.001), remained independent predictors.

Conclusions: While several factors were related to HR-HPV persistence in univariate analysis and when adjusted for age and follow-up time as confounders, the only independent predictors in the multivariate regression model were follow-up time and HR-HPV positive PCR. Clearly more data are needed on type-specific persistence and HPV integration as its predictors.