To analyze the effect of highly active anti-retroviral therapy on cancer incidence and mean age at cancer diagnosis in HIV-infected persons; on progression to AIDS related to cancer; and risks of cancer, pre-and post-HAART. Methods Records on New Jersey HIV-infected residents were used for a retrospective cohort study of HIV and cancer. Lung, colorectal, prostate, Hodgkin lymphoma, and anal cancer were studied. Incidence density pre-and post-HAART was tested for change in Standardized Incidence

more »

... Ratio (SIRs). Relative risk was approximated by the ratio of the SIRs from the two periods. Difference in mean age at diagnosis pre-and post-HAART for each type was tested by a Student's t. Risks of developing AIDS, pre-and post-HAART, in HIV-infected persons with and without a cancer diagnosis were evaluated through proportional hazards iii regression. Cox Proportional Regression was used to assess the hazard ratios associated with cancer after HIV seroconversion in the two periods. Results Cumulative incidence rates increased substantially for prostate cancer, and slightly for anal cancer and Hodgkin lymphoma pre-to post-HAART, but not for lung and colorectal cancer. Cumulative incidence rates for HIV-infected persons observed only pre-HAART were lower when compared to those HIV-positive only post-HAART for all five studied cancer types. SIRs were significant for all but pre-HAART colorectal cancer. Relative risks post-HAART compared to pre-HAART for each of the cancers, other than Hodgkin lymphoma, were significantly low. Age-specific incidence rates post-HAART were higher in the oldest age groups for colorectal and prostate cancer in addition to Hodgkin disease, but decreased in every age group for lung cancer. Mean age at cancer diagnosis was significantly higher in the post-HAART period for lung and colorectal cancer and Hodgkin lymphoma, but unchanged for prostate cancer. The regression analysis indicated an increased risk of AIDS development if a cancer was diagnosed after HIV seroconversion, but before AIDS, although the risk decreased post-HAART. The hazard ratio for cancer development was not significantly different pre-and post-HAART. Conclusions HAART availability and an increase in survival have significantly changed trends in HIV infection and associated morbidity, particularly for cancer in HIV-infected persons. iv ACKNOWLEDGEMENTS