Cancer incidence and cancer-attributable mortality among persons with AIDS in the United States: 1980--2006

The aims of this dissertation were to: (1) determine cancer risk among persons with long-standing AIDS in years 3-5 and years 6-10 after AIDS onset, (2) to evaluate the impact of HAART on cancer incidence in years 3-10 after AIDS onset, (3) to quantify the cumulative incidence of AIDS-defining cancer and non-AIDS-defining cancer, controlling for trends in mortality, and (4) to determine the fraction of deaths among persons with AIDS attributable to cancer.;Data from the population-based U.S. HIV/AIDS Cancer Match Study (HACM) were used to address the aims of this dissertation. Records of persons with HIV/AIDS in surveillance registries from 9 states and 6 metropolitan areas (diagnosed during 1980-2008) were linked to corresponding cancer registry records using a probabilistic matching algorithm. Subsequent to the match, all identifying information was removed. For Aims 1 and 2, we constructed a cohort of 263,254 adults and adolescents with AIDS (diagnosed during 1980-2004) and evaluated incident cancers occurring during years 3-5 and 6-10 after AIDS onset. Standardized incidence ratios (SIRs) assessed risks relative to the general population. Rate ratios (RRs) derived from Poisson regression compared cancer incidence before and after 1996 to assess the impact of HAART. For Aims 3 and 4, we constructed a cohort of 372,364 adults and adolescents with AIDS who were alive, cancer-free, and under follow-up for cancer at the start of the fourth month after AIDS onset. We used competing risk methods to determine cumulative incidence of cancer (AIDS-defining cancers [ADCs] and non-AIDS-defining cancers [NADCs]) and Cox regression to estimate cancer-attributable mortality across 3 calendar periods (AIDS onset in 1980-1989, 1990-1995, and 1996-2006).;Aims 1 and 2 demonstrated risks of ADCs (Kaposi sarcoma, non-Hodgkin lymphoma, cervical cancer) were significantly elevated during 3-5 and 6-10 years after AIDS, and incidence of Kaposi sarcoma and non-Hodgkin lymphoma declined significantly between the pre-HAART (1990-1995) and HAART era (1996-2006). Other cancers with elevated risks in the 3-5 and 6-10 year periods, respectively, were cancers of the oral cavity/pharynx (SIR 1.9 95%CI 1.6-2.1, and SIR 1.8 95%CI 1.5-2.1) and anus (SIR 27 95%CI 24-31, and SIR 40 95%CI 35-45), and Hodgkin lymphoma (SIR 9.1 95%CI 7.8-11, and SIR 12 95%CI 9.7-14). Between 1990-1995 and 1996-2006, incidence increased for anal cancer (RR 2.9 95%CI 2.1-4.0) and Hodgkin lymphoma (RR 2.0 95%CI 1.3-2.9). Aims 3 and 4 demonstrated cumulative incidence of ADCs declined across AIDS calendar periods (from 8.7% among persons diagnosed with AIDS during 1980-1989 to 6.4% among persons diagnosed with AIDS during 1990-1995 to 2.1% among persons diagnosed with AIDS during 1996-2006). Cumulative incidence of NADCs increased from 0.86% in 1980-1989 to 1.1% in 1990-1995 with little change thereafter (1.0%, 1996-2006). However, the cumulative incidence of some site-specific NADCs (anal cancer, Hodgkin lymphoma, liver cancer and lung cancer) increased. Among those with AIDS and cancer, cancer-attributable mortality increased to 88.3% (ADC) and 87.1% (NADC) during 1996-2006, and population-attributable NADC mortality increased to 2.3% (1996-2006). Population-attributable ADC mortality decreased from 6.3% (1990-1995) to 3.9% (1996-2006).;Among people who survived an AIDS diagnosis for several years or more, we observed continuing risks of ADCs and elevated long-term risks for selected NADCs, notably anal cancer and Hodgkin lymphoma. We also noted dramatically declining incidence of ADCs and increases in some NADCs, while controlling for temporal trends in mortality using competing risk methods. Among people with AIDS who develop cancer, their malignancy is the predominant cause of death, pointing to the need for more effective cancer treatment in this population. Further, NADCs account for a growing fraction of all deaths among persons diagnosed with AIDS in the HAART era. Continued monitoring of long-term cancer risk among persons with AIDS is warranted and should be extended to include persons with HIV infection alone. As HIV infection is increasingly considered with chronic disease management paradigms, greater attention should be focused on cancer screening and prevention strategies among person with HIV/AIDS. (Abstract shortened by UMI.)...