Risk of cancer among women with AIDS in New York City.

To evaluate the risk of cancer among women with AIDS in New York City (NYC), we compared the cancer experience of AIDS-infected women in NYC with that of the general population of women in NYC by matching the population-based New York State Cancer Registry with the New York City AIDS Registry. A probabilistic algorithm was used to match names, birth dates, and, where available, Social Security numbers between 15,146 women with AIDS and 232,902 women with cancer. Standardized incidence ratios (SIR) were calculated as the ratio of observed to expected cancer cases in the population of NYC women matched for age, race, and calendar period of cancer diagnosis. Period-specific relative risks (RR) of cancer prevalence prior to AIDS, and incidence at or after AIDS were calculated to determine which cancers increased in proximity to an AIDS diagnosis, a surrogate marker of increasing immunodeficiency. Analysis was limited to women between the ages of 15 to 69 who were diagnosed with AIDS between 1981 and 1994. Among 15,146 women diagnosed with AIDS, we found 1,194 matches with the Cancer Registry. For cancers included in the 1993 AIDS case definition, the SIR was 178.49 for Kaposi's sarcoma, 48.97 for non-Hodgkin's lymphoma, and 9.20 for invasive cervical cancer. The overall SIR for all non-AIDS-defining cancers was 2.20. Among non-AIDS-defining cancers, elevated SIRs were found for cancers of the lung (7.95), esophagus (7.69), multiple myeloma (7.37), oral cavity and pharynx (6.55), Hodgkin's disease (5.65), leukemias (4.52), and rectal/anal cancers (3.23). Statistically significant increases in period-specific risks were found for all non-AIDS-defining cancers combined, but not for individual cancers. Dual screening by two registries and unknown behavioral factors complicate the ascertainment of cancer risk. Our results show significantly elevated risks for several non-AIDS-defining cancers; these results are consistent with other studies of cancers among persons with AIDS. Extension of the time period of analysis is required to test for the effects of new anti-viral treatments and their association with cancer development among HIV-infected women.