HIV AND CANCER IN KAMPALA, UGANDA.

2 In a study of the relationship between HIV and cancer, in Kampala, Uganda, subjects with incident cancer are recruited from the main hospitals and submitted to an HIV test. Between August 1994 and October 1997, 1749 individuals with cancer were tested for HIV. The HIV seroprevalence in all cancers excluding Kaposi's sarcoma, non-Hodgkin's lymphoma and squamous cell carcinoma of the conjunctiva (which are already known to be HIV-associated) was 16%, in line with independent seroprevalence data from Kampala. These were used as a comparison group when calculating relative risks for specific cancer sites. As expected, there was a significantly increased risk of Kaposi's sarcoma in relation to HIV infection (OR 19.2; 13.2-27.8). This excess risk is lower than has been found in the USA, perhaps reflecting the higher incidence of HIV seronegative "endemic" Kaposi's sarcoma in Uganda, which was relatively common there even before the advent of AIDS. There was also a lower than expected excess risk of non-Hodgkin's lymphoma in HIV seropositive subjects (OR 1.9; 1.0-3.5). Only 1/6 cases of Burkitt's lymphoma was HIV seropositive (OR 0.4; 0.0-4.1), suggesting that any excess risk is unlikely to be as large as in the West. Squamous cell carcinoma of the conjunctiva was associated with HIV infection (OR 6.9; 4.2-11.6), consistent with results from elsewhere in Africa. HIV infection does not appear to greatly increase the risk of hepatocellular carcinoma (OR 0.8; 0.4-1.7), or cervical cancer (OR 1.2; 0.8-1.8), but is associated with an increased risk of penile cancer (OR 3.7; 1.3-10.1). This has not been noticed before and bears further investigation. The only other malignancies that show a significant association with HIV infection were oro-pharyngeal cancers (OR 2.4; 1.1-5.3), also an unexpected finding.