In this large cohort of Zimbabwean women, we found consistently increased risk of HIV acquisition associated with cervico-vaginal HPV infection, and specific HPV types (HPV 31, 58 and 70) were implicated in the association. These results suggest a previously unrecognized role of HPV in HIV acquisition among women. An obvious explanation for the observed association is that it is attributable to a partner with both infections, and we cannot definitively rule out the effect of sexual exposure to both infections as the explanation for the association. However, demographic information collected at enrollment and at each quarterly visit allowed us to control for time-dependent variables such as new sexual partners, HSV-2 serostatus, new STIs, or high-risk behavior by the participant or her regular partner. In addition, the results from our sensitivity analysis suggest that HPV infection preceded HIV acquisition.
Most HPV infections are cleared by innate and adaptive immune mechanisms 
, making it is biologically plausible that the local immune response elicited by HPV infection predisposes to HIV acquisition, and that women in the process of clearing an HPV infection might be at increased risk. This possibility is supported by our finding that non-persistent HPV infections were associated with increased risk of HIV acquisition whereas persistent infections were not. HPV infection does not cause epithelial ulcerative lesions but rather is associated with hyperproliferative changes such as warts, cervical intraepithelial neoplasia or cancer; these lesions are known to be infiltrated by HIV target cells such as lymphocytes and macrophages 
, . Persistent HPV infections are associated with an increased risk of developing precancerous lesions, but in our study were not associated with an increased risk of HIV acquisition. These results suggest that clearance of HPV infection rather than the dysplastic effect of HPV on epithelial cells confers HIV risk. We do not known whether the women in this study had HPV-associated lesions at the time of HIV acquisition, and although we found no significant association between abnormal cytology (which detects HPV-associated lesions) and risk of HIV infection, this result may have been limited by lack of power. Additional studies incorporating frequent cytological and colposcopic evaluations are required to definitively establish the role of HPV-related lesions in HIV transmission. The observed association between HPV infection and HIV acquisition might also be due to unmeasured confounding variables such as social networks; although we adjusted for multiple variables that measure risk of HIV through sexual activity, this report cannot definitively establish the exact role of HPV infection in risk of HIV acquisition.
Our finding of HPV type-specific associations (HPV 31, 58 and 70) raises the possibility that biological processes might underlie the association with HIV risk. HPV 70 is a non-oncogenic HPV type in clade A7 whereas HPV 31 and 58 are oncogenic HPV types in clade A9. HPV16, another member of clade A9, is the most common HPV type in the African continent 
; it was second in prevalence after HPV 58 in the Zimbabwean women in our study (4.7%), but it did not confer significantly increased risk of HIV acquisition in any of our analyses. The reason for the HIV risk association with a subset of HPV types is unknown but may reflect type-specific differences in the host immune response to HPV. The relatively high prevalence of HPV 58 and HPV 70 in this cohort, and the observed association of these types with HIV acquisition, suggest that the contribution of HPV to HIV acquisition risk may vary in different populations depending on the distribution of HPV types and host genetic polymorphisms. The currently available HPV vaccines target HPV 6, 11, 16 and 18, which are not the HPV types identified with HIV risk in our study (HPV 31, 58 and 70). Further investigation into host and HPV type-specific differences in the immune/inflammatory responses to HPV infection will help to elucidate possible biological mechanisms by which HPV infection could contribute to HIV transmission.
The strengths of our study are the large sample size, the high retention rate, and the abundance of data about HPV status and risk factors collected at frequent intervals over time. A limitation of this study is that the commonly used definition of HPV infection cannot differentiate between the presence of biologically active HPV versus
the detection of HPV DNA due to deposition in the genital tract during a recent sexual encounter. For this reason and due to the observational study design, these results cannot establish causation between HPV infection and HIV acquisition, but they suggest that HPV is a potentially significant risk factor for HIV acquisition in women. To date, little is known about the association between HPV and HIV transmission. In a recent study in men who have sex with men, anal infection with 2 or more HPV types significantly increased the risk of HIV infection (HR 3.5, 95% CI
. Given the high prevalence of anogenital HPV infection, further investigation of the role of HPV in HIV acquisition is warranted. A better understanding of the factors underlying the HPV-HIV association will improve our understanding of the pathogenesis of HIV transmission.