Cumulative Incidence of Cancer Among Persons With HIV in North America: A Cohort Study

Abstract

Background: Cancer is increasingly common among persons with HIV.

Objective: To examine calendar trends in cumulative cancer incidence and hazard rate by HIV status.

Design: Cohort study.

Setting: North American AIDS Cohort Collaboration on Research and Design during 1996 to 2009.

Participants: 86 620 persons with HIV and 196 987 uninfected adults.

Measurements: Cancer type-specific cumulative incidence by age 75 years and calendar trends in cumulative incidence and hazard rates, each by HIV status.

Results: Cumulative incidences of cancer by age 75 years for persons with and without HIV, respectively, were as follows: Kaposi sarcoma, 4.4% and 0.01%; non-Hodgkin lymphoma, 4.5% and 0.7%; lung cancer, 3.4% and 2.8%; anal cancer, 1.5% and 0.05%; colorectal cancer, 1.0% and 1.5%; liver cancer, 1.1% and 0.4%; Hodgkin lymphoma, 0.9% and 0.09%; melanoma, 0.5% and 0.6%; and oral cavity/pharyngeal cancer, 0.8% and 0.8%. Among persons with HIV, calendar trends in cumulative incidence and hazard rate decreased for Kaposi sarcoma and non-Hodgkin lymphoma. For anal, colorectal, and liver cancer, increasing cumulative incidence, but not hazard rate trends, were due to the decreasing mortality rate trend (-9% per year), allowing greater opportunity to be diagnosed. Despite decreasing hazard rate trends for lung cancer, Hodgkin lymphoma, and melanoma, cumulative incidence trends were not seen because of the compensating effect of the declining mortality rate.

Limitation: Secular trends in screening, smoking, and viral co-infections were not evaluated.

Conclusion: Cumulative cancer incidence by age 75 years, approximating lifetime risk in persons with HIV, may have clinical utility in this population. The high cumulative incidences by age 75 years for Kaposi sarcoma, non-Hodgkin lymphoma, and lung cancer support early and sustained antiretroviral therapy and smoking cessation.

Figure 1. Cumulative cancer incidence by calendar era and cancer type for HIV-infected subjects with age as the time scale, NA-ACCORD, 1996–2009

KS, Kaposi sarcoma; NHL, Non-Hodgkin’s lymphoma; HL, Hodgkin lymphoma. Cumulative incidence (i.e., cancer risk) curves for calendar eras 1996–1999 (light gray dash-dot line), 2000–2004 (dark gray dash line), and 2005–2009 (black solid line), were obtained using non-parametric estimators for competing risk events. Numbers associated with curves represent cumulative incidence (%) by age 75 years; *P<0.05 compared with reference (1995–1999 era); ** P<0.01 compared with reference (1995–1999 era).

Figure 2. Cumulative cancer incidence by calendar era and cancer type for uninfected subjects with age as the time scale, NA-ACCORD, 1996–2009

KS, Kaposi sarcoma; NHL, Non-Hodgkin’s lymphoma; HL, Hodgkin lymphoma. Cumulative incidence (i.e., cancer risk) curves for calendar eras 1996–1999 (light gray dash-dot line), 2000–2004 (dark gray dash line), and 2005–2009 (black solid line), were obtained using non-parametric estimators for competing risk events. Numbers associated with curves represent cumulative incidence (%) by age 75 years; *P<0.05 compared with reference (1995–1999 era); ** P<0.01 compared with reference (1995–1999 era).