The prevention of cervical cancer in HIV-infected women

HIV-infected women have a substantially increased risk of human papillomavirus (HPV) infections compared to HIV-uninfected women [1,2]. The reasons for this finding have been debated for some time. HPV infection is predominantly acquired through heterosexual intercourse, as, in the vast majority of women worldwide, is HIV infection. A confounding effect of sexual behaviour was, therefore, evoked as a possible explanation of the excess of HPV infections and HPV-related anogenital cancers in HIV-infected women [3]. However, certain features of HPV infection in HIV-infected women cannot be explained by increased HPV exposure alone. For example, both the prevalence and the incidence of HPV in HIV-infected women are inversely associated with CD4+ cell level [1]. Moreover, new HPV infections appear in the absence of recent sexual activity much more often in HIV-infected than HIV-uninfected women, most probably a result of reactivation of latent infections [1,4]. Finally, HIV-infected women are more likely than HIV-uninfected women to harbour multiple-type infections [1,2] and to develop high-grade cervical lesions from oncogenic types other than HPV16 [1,2]. The relative risk for HPV infection among HIV-infected compared to HIV-uninfected women is consistent across different populations (between 2 and 4, depending on the degree of immunosuppression [1,2]). The relative risk for cervical cancer, however, varies enormously (from less than 2 to over 20), as it is affected by premature death from other causes or early detection and treatment of cervical precancerous lesions [5,6]. Unfortunately, cervical cancer prevention in HIV-infected women is lagging behind even in the highest-resource countries [5]. For instance, a combined analysis of 13 North American cohort studies of HIV-infected women (n = 16 467) showed that between 1995 and 2007, cervical cancer incidence was 10-fold higher among HIV-infected women than the general female population, despite the widespread use of HAART and substantial screening efforts [7]. This raises the important question of whether the vast progresses in the understanding of cervical cancer screening in the general female population [8] can also be applied to HIV-infected women. There is no doubt about the equal importance of high coverage and follow-up of screening-positive women [8]. More recently, however, large randomized clinical trials from high-resource and low-resource countries have shown that in the general female population HPV test-based screening allows for earlier detection of precancerous lesions (cervical intraepithelial neoplasia, CIN, 2 and 3). Hence, it confers longer protection against CIN2 or worse (+) and cervical cancer than does cytology [9–13] or visual inspection with acetic acid (VIA) [12]. In addition, HPV testing is less dependent on high standards of quality assurance than cytology or VIA [8]. No recommendations exist for the use of HPV testing for primary screening or triage in HIV-infected women [14], and concerns have been raised about its low specificity [15,16]. But support for the use of HPV testing to screen HIV-infected women has now been provided for the first time in an important article in the current issue of AIDS[17]. The article shows the outcome of a large randomized trial carried out in South Africa, which included 6553 women aged 35–65 years, of whom 956 were infected with HIV. One of two different tests (Hybrid Capture 2 or VIA) was combined with immediate cryotherapy of screening-positive women (screen-and-treat) and the two different testing groups were compared with a control group in which evaluation or treatment was delayed for 6 months. The endpoints were very robust ones: CIN2+ or CIN3+ detected by colposcopy and biopsies at month 6 and, in a subset of study women, month 12, 24, and 36. Screen-and-treat using HPV testing was demonstrated to be as feasible, safe, and efficacious in HIV-infected women as it was in HIV-uninfected women. As expected, HPV test specificity was lower, but sensitivity and positive and negative predictive values were not compromised by the presence of HIV infection. The number of CIN2+ prevented per 100 women screened was actually greater among HIV-infected women (11.9) than among HIV-uninfected (3.1) women. Screen-and-treat based on VIA was substantially less beneficial than that based on HPV testing, mainly on account of the low sensitivity of visual methods even in skilled hands. On the negative side, after cryotherapy, HIV-infected women continued to have many more HPV infections [18] and CIN1 (i.e., the histological manifestation of HPV infection) [17] than HIV-uninfected women. The findings on HPV testing reported by Kuhn et al.[17] may also be indirectly relevant to HIV-infected women outside sub-Saharan Africa, regardless of whether screen-and-treat is used. High HPV prevalence and the consequent loss of specificity for CIN2+ may not be a good enough reason to forgo using the best screening test (i.e., HPV testing) if the positive predictive value for CIN2+ is not lowered [17]. In fact, HPV-positive women who are also HIV-infected have been shown to have a similar [17], or even higher [19], probability of developing CIN2+ than HIV-uninfected women. A larger percentage of screening-positive findings and of subsequent lesion work-up or treatment may be an unavoidable price to pay to prevent cervical cancer in HIV-infected women. However, a few important issues need more research, most notably the short-term influence of cryotherapy and other cervical treatments on the probability of transmitting HIV to a male partner, and the best management of HPV-positive women who are also infected with HIV. The way to distinguish many transient HPV infections from long-duration cancer-inducing infections, however, is currently the most important priority to improve HPV test-based screening regardless of HIV status [20].