Higher rates of non-AIDS cancers in HIV positive people
Nathan Geffen, TAC
Michael Silverberg presented a case-control retrospective analysis from the Kaiser Permanante cohort of HIV-positive and HIV negative patients, to determine differences in non-AIDS cancer rates post HAART (1996-2007). 
They found significantly higher non-AIDS infection-related cancers (anal, Hodgkins, liver, oral cavity/pharynx) in the HIV-positive group, as reported in many previous studies.
The analysis compared over 20,300 HIV-positive patients, matched for age and sex, with ten HIV-negative cases each (i.e. over 203,000 controls), giving well over a million years of follow-up. The mean follow-up period was 4.2 and 5 years in the HIV-positive and HIV-negative groups respectively. Mean age was 40 years. This was a predominantly white MSM male (90%) cohort. Patients were followed until cancer, death, the end date of the study or they left the health provider. Time analyses looked at three four-year periods: 1996-99, 2000-3 and 2004-7.
Non-AIDS defining cancers were diagnosed in 3% of the HIV-positive and 2% of the HIV negative groups. Rates for infection-related non-AIDS-defining cancers (per 10,000 person-years) were 29.7 for HIV positive and 4.4 for HIV-negative patients. Infection-related cancers accounted for 46% and 13% of the cancers in HIV-positive and HIV negative subjects respectively. This difference was significant for any infection-related cancer [RR: 7.4; 95%CI 6.4 8.5], or each taken individually. It was particularly high for anal cancer [RR: 80; 95%CI 50.2 126.4], Hodgkins lymphoma (RR 17.4; p <0.001), head and neck (RR 2.1; p <0.001), and gynecologic (RR 2.9; p = 0.001).
Interestingly, the incidence rates for any infection-related cancer came down 4% per year in HIV-positive subjects and went up by 4% per year in HIV-negative patients (p<0.001, adjusted for age and sex). This difference between the two groups was also significant for each infection-related cancer analysed separately, again with anal cancer having the largest per annum decrease (6% decline versus 13% increase, p<0.001). The 4% per year decline in any infection-related cancer over time in HIV positive subjects was significant (p=0.003), but this was not significant when each cancer was considered separately.
The risk of any non-infection related cancer (lung, melanoma, kidney, hematologic, colorectal, prostate) was also higher in the HIV-positive group [RR: 1.2; 95%CI 1.11.4], but the differences were less pronounced. Analysed separately, this was only significant for lung, melanoma and kidney cancers. Except for lung cancer, there were no statistically significant time-related differences.
Silverberg pointed out the studys strengths: its large size, that the two comparative groups were drawn from the same population and that the cancer data was obtained from comprehensive registries. But he also noted the considerable limitations: no consideration of smoking, CD4 count or cancer screening practices. Also the study has limited data on women, lower-income (uninsured) patients and non-white patients.
While the increased relative risk was high, the absolute risk of cancer was small during the follow-up period for both groups. Of course, with a longer follow-up period that would undoubtedly rise given that cancer is a common cause of mortality, irrespective of HIV status.
Despite the studys limitations, the high correllation between being HIV positive and increased risk of cancer in this cohort indicates that regular cancer screening may be particularly important for these patients.
- Silverberg M. Infection-related Non-AIDS defining cancer risk in HIV-infected and -uninfected persons. 16th CROI, Montreal, 2009. Oral abstract 30.