LEGAME TRA L¿IMMUNODEFICIENZA HIV-CORRELATA E L¿INSORGENZA DI TUMORI: ASPETTI DI METODOLOGIA STATISTICA

The aim of the present Ph.D. program was exploring methodological issues arising in studies of cancer incidence, relative risk, and survival in patients with HIV/AIDS (PHA). Aspects related to the first two objectives were explored in the first two years. This thesis described a record-linkage study conducted between the national Italian AIDS Registry and 24 Italian cancer registries to estimate survival after a cancer diagnosis in PHA. More than 2600 cancer cases diagnosed between 1986 and 2005 were included. Survival in PHA was compared with that reported in patients without AIDS using, as comparison group, patients matched for site (1:1 for Kaposi Sarcoma, 1:2 for non-Hodgkin lymphoma, 1:5 for other cancers), sex, age, period of diagnosis, and area of residence. Overall survival and death hazard ratios (HR) compared survival in PWA with cancer to that in cancer patients without AIDS have been calculated. Overall, the 3-year survival rate of PHA with cancer increased from 16% in 1986-1995 to 41% in 1996-2005 period, after the widespread use of antiretroviral therapy (cART). In this period, HR remained higher in PHA than in persons without AIDS (3.0, 95% confidence interval [CI]: 2.7–3.4), in particular for cancer with good prognosis, e.g., Hodgkin lymphomas (HR=8.6), non-melanoma skin cancer (H=5.0), and anal cancer (HR=4.0). A sensitivity analysis was performed to evaluate the impact on survival and on HR of different study designs and comparison groups.