Cancer Causes 10% of Deaths Among People With HIV in U.S., Canada

One in 10 deaths among HIV-positive North Americans from 1995 to 2009 can be attributed to cancer, according to a National Cancer Institute (NCI) analysis of 46,956 people taking antiretroviral therapy (ART). The fraction of deaths attributable to cancer rose over the study period.

People with HIV have a heightened risk for AIDS-associated cancers and some non-AIDS-associated cancers because of HIV-related immune abnormalities, infection with cancer-causing viruses and lifestyle risks such as smoking. Research indicates worse survival with cancer in people with HIV than in HIV-negative individuals.

With collaborators in the North American AIDS Cohort Collaboration on Research and Design (NA-ACCORD), NCI investigators analyzed the population-attributable fraction (PAF) of deaths due to cancer in adults taking ART.

NA-ACCORD includes more than 20 cohorts of people in care for HIV in the U.S. and Canada. The NCI analysis involved 13 cohorts that had data on clinical characteristics, cancer incidence and mortality. Follow-up ranged from January 1995 through December 2009. The primary analysis excluded people with cancer before follow-up began.

The researchers used Cox proportional hazards models to determine hazard ratios (HR) for associations between a cancer diagnosis and overall mortality. These models were adjusted for age, sex, race, HIV risk group, year, cohort, CD4 count and viral load. The investigators calculated PAF as Pd x (HR - 1)/HR, where Pd is the proportion of people who died after a cancer diagnosis. They calculated cancer-attributable mortality as PAF x (number of deaths)/(person-time of follow-up).

The study focused on 46,956 antiretroviral-treated adults with a median age of 42.5 years, 83.3% of them men, 45.6% white and 40.7% black. Three-quarters of the group (75.5%) had ever smoked, 30.3% had a CD4 count below 200 cells/mm3 and 40.8% had a viral load below 500 copies/mL.

During an average 5.7 years of follow-up, 8,956 people died, yielding an overall mortality of 3,352 per 100,000 person-years. Among 1997 people with a first cancer diagnosis, 1069 (53.5%) died, and cancer accounted for 11.9% of all deaths (Pd). In the adjusted analysis, cancer raised the risk of dying more than five times (HR 5.54). Adjusted PAF for overall cancer-attributable mortality was 9.8% (95% confidence interval 9.1% to 10.5%). Cancer-attributable mortality came to 327 per 100,000 person-years, much higher than the estimated 186 per 100,000 in the general U.S. population.

AIDS-associated cancers explained 2.6% of deaths and non-AIDS-associated cancers explained 7.1%. Individual cancers with the highest PAFs were lung cancer (2.3%), non-Hodgkin lymphoma (2.0%), liver cancer (0.9%), Kaposi sarcoma (0.5%) and anal cancer (0.4%). Men had higher cancer-attributable mortality than women (10.2% versus 7.2%), largely because of non-AIDS-associated cancers (7.5% versus 4.9%).

Cancer-attributable mortality rose with age to reach 13.9% in people 55 years old or older, again largely because of non-AIDS-associated cancers (2.1% among people under age 40 years and 12.5% among people age 55 and older). People with a CD4 count below 100 cells/mm3 had the highest cancer-attributable mortality (1,068 per 100,000 person-years), and that rate fell with increasing CD4 counts (96 per 100,000 person-years with ≥500 cells/mm3). Cancer PAF rose from 7.7% before 2001, to 8.1% by mid 2003, to 10.5% by 2005 and to 12.1% after 2005. For AIDS-associated cancers, PAF dropped from 3.4% before 2001 to 1.9% after 2005, while for non-AIDS-associated cancers PAF rose from 4.2% to 10.1% across those periods.

The researchers conclude that about 10% of deaths among people prescribed ART from 1995 to 2009 can be attributed to cancer, and this rate rose over time. They note that the PAFs reflect both high cancer incidence in people with HIV and high subsequent mortality. The authors predict that deaths caused by non-AIDS-associated cancer "will likely grow in importance as AIDS mortality declines" and HIV-positive populations age.

The NCI team believes their findings underline opportunities for cutting cancer mortality by (1) lowering cancer incidence "by facilitating adherence to ART, enabling smoking cessation, and treating [hepatitis B and C] infection," (2) screening for lung, liver and anal cancers in high-risk populations and (3) ensuring access to "timely and effective cancer treatment" for people with HIV infection.