Background:  The extent to which HAART affects human papillomavirus (HPV) infection and clearance in HIV-infected women is not well understood. Because most HPV infections are transient and can recur, infrequent observations (6 to 12 months), few visits in studies (3 to 5 years), and missed visits pose challenges in HPV studies. To address these, a novel statistical approach that uses incomplete data efficiently to describe HPV/HIV disease process was applied to published ACTG A5029 trial:  an observational study with 4 scheduled visits (0, 24, 48, 96 weeks) in 147 ART-naïve HIV-infected women initiating HAART (≥3 drugs with PI or NNRTI, or 3-nucleoside regimen with abacavir).

Methods:  Markov models were applied to describe transition rates over time since HAART initiation, where states were defined by presence or lack of high risk HPV (any type:  16, 18, 26, 31, 33, 35, 39, 45, 51, 52, 53, 56, 58, 59, 66, 68, 73, 82) and current CD4 (≤350 or >350 cells/mm³), and separately to describe high risk HPV and current plasma HIV RNA viral load (viral load, ≤400 or >400 copies/mL). Rates of HPV acquisition (initial or recurrent) and clearance over time based on current (time-varying) CD4 and viral load were assessed.

Results:  Median age, baseline CD4 and viral load were 35 years, 238 cells/mm³, and 13,894 copies/mL. high risk HPV was common at baseline (62%). As expected, HAART improved CD4 and viral load (72% CD4>350 and 74% viral load <400 at week 96). Higher CD4 was significantly associated with HPV clearance (P=0.001, hazard ratio=3.93). Higher viral load was associated with higher rate of HPV acquisition (P=0.068, hazard ratio=4.67; or P=0.016, hazard ratio=4.84, depending on the specific model used). Other comparisons were not statistically significant; there was no evidence that HPV affects HIV disease. Estimated state probabilities suggest that a woman starting HPV^– and CD4 ≤350 has the following chances of being in the other states 1 year after HAART initiation:  19% in HPV⁺ and CD4 ≤350, 36% in HPV^– and CD4 >350, and 13% in HPV⁺ and CD4 >350. They stabilize in 2.5 years to around 15%, 50%, and 25%, respectively. Probabilities stabilize in about 2 years for a woman starting HPV^– and viral load >400 to 10% (HPV⁺, viral load >400), 50% (HPV^–, viral load ≤400) and 30% (HPV⁺, viral load ≤400). Assessed model fits comparing observed and expected transition counts were adequate.

Conclusions:  Higher CD4 was associated with HPV clearance, and higher viral load was associated with HPV acquisition. Use of HAART to improve CD4 and decrease viral load may influence the control of HPV infection.