Non-AIDS-defining malignancies in HIV-infected persons: etiologic puzzles, epidemiologic perils, prevention opportunities

Abstract

Non-AIDS-defining malignancies have come to represent a growing fraction of the overall cancer burden in HIV-infected people, as improvements in HIV therapy prolong survival and reduce the incidence of AIDS-associated cancers. This review focuses on five non-AIDS-defining malignancies for which HIV-infected persons have an elevated risk, for which risk is substantial or increasing over time, and for which HIV infection may play an etiologic role. Among HIV-infected persons, lung cancer risk is high, in part due to frequent tobacco use in this population. Risks of anal cancer and liver cancer are also elevated, related to the high prevalence of infections with human papillomavirus and hepatitis B and C viruses. In addition, risk is elevated for Hodgkin lymphoma and several rare skin cancers, including Merkel cell carcinoma and sebaceous carcinoma. For anal cancer and Hodgkin lymphoma, it is particularly concerning that incidence in HIV-infected persons has risen in recent years, when highly active antiretroviral therapy has been available. Accumulating evidence supports the possibility that the high prevalence of known carcinogenic exposures (e.g., tobacco) and infections with oncogenic viruses does not completely explain the occurrence of these cancers. Indeed, HIV may act to increase the risk for each of these five non-AIDS-defining malignancies, although the mechanisms may vary, including immunosuppression, immune reconstitution, and chronic inflammation. These non-AIDS-defining cancers also present important opportunities for prevention (e.g., smoking cessation), screening (e.g., periodic anal Pap smear screening), and early detection.

Figure 1

Cancer incidence among people with AIDS in the U.S. (1984-2002). Incidence is shown as a function of calendar year of AIDS onset for Kaposi sarcoma (KS), non-Hodgkin lymphoma (NHL), cervical cancer, and non-AIDS-defining cancers. Incidence estimates for each cancer are stacked on top of each other, to depict the proportion of total cancer incidence contributed by each cancer type. Data pertain to the two-year period following AIDS onset. Data are from the HIV/AIDS Cancer Match Study. The figure is reproduced with permission from Engels et al. (1).

Figure 2

Hodgkin lymphoma risk among people with AIDS in the U.S. Risk is shown as a standardized incidence ratio (SIR), measuring risk relative to that in the general population. Panel A depicts SIRs as a function of year of AIDS onset. The smooth curve corresponds to a significant linear trend of increasing Hodgkin lymphoma risk during 1990-2002; the data points corresponding to AIDS onset in 2000-2002 are based on sparse numbers of people under follow-up. Panel B depicts SIRs as a function of CD4 count at AIDS onset (cells/mm³). The smooth curve corresponds to a model with a significant quadratic term, indicating maximum Hodgkin lymphoma risk at a CD4 count around 225-249 cells/mm³. Data are from the HIV/AIDS Cancer Match Study. Panels are similar to figures presented in Engels et al. (1) and Biggar et al. (32).