CROI 2015 Program and Abstracts

Abstract Listing

Oral Abstracts

Conclusions: ART reduces the risk of developing cancer in HIV-infected children in South Africa. Early linkage to care and early start of ART may help to further reduce the burden of cancer in these children. 725 High Cancer Risk Among the HIV-Infected Elderly in the United States Elizabeth L. Yanik ; Hormuzd A. Katki; Eric A. Engels National Cancer Institute (NCI), Rockville, MD, US Background: HIV-infected people have higher risk of many cancers compared to HIV-uninfected people, but it is unclear if the magnitude of this elevated risk is consistent across age groups. As the proportion of HIV-infected people over age 65 is increasing over time and the elderly population is known to have high cancer risk, it is important to understand the relationship between HIV and cancer in this age group. Methods: We conducted a case-cohort study that included a 5% sample of Medicare enrollees and all cancer cases ≥ 65 years of age identified through the Surveillance, Epidemiology, and End Results cancer registries. Non-melanoma skin cancers were not captured. HIV infection was defined through Medicare diagnosis claims. Weighted Cox regression was used to estimate associations between HIV and cancer incidence adjusting for age, race, sex, and calendar year. The absolute risk of cancer over time was calculated accounting for the competing risk of death. Results: Among 469,954 people in the 5%Medicare sample, 0.08% had an HIV diagnosis. In total, 835,450 cancer cases were identified in cancer registries. Among HIV-infected people, lung and prostate cancers were most common (N=111 each), followed by non-Hodgkin lymphoma (NHL) (N=57). HIV was strongly associated with incidence of Kaposi sarcoma, anal cancer and Hodgkin lymphoma (hazard ratios of 104, 30, and 12, respectively, Table 1). HIV was also associated with incidence of liver cancer, NHL, and lung cancer, but elevations in risk were lower (hazard ratios of 5, 3 and 2, respectively). Among NHL subtypes, HIV was associated with diffuse large B-cell lymphoma and Burkitt lymphoma, but no association was found with other specified NHL subtypes (which comprised 60% of cases in uninfected people). HIV was associated with lower prostate cancer incidence. Over a 1-year period, 2.5% of the HIV-infected elderly were diagnosed with cancer; by 5 years, this proportion increased to 10.2%.

Oral Abstracts

Conclusions: HIV infection in the elderly is associated with higher risk for many cancers identified as HIV-associated in younger populations. The relative elevation in NHL incidence is notably lower, but this reflects the high frequency in elderly adults of NHL subtypes less strongly associated with HIV. Given the increased risk associated with both aging and HIV, the elderly HIV-infected population has a sizeable absolute risk of cancer, highlighting the need for cancer prevention and screening efforts in this group.

726 Smoking Outweighs HIV-Related Risk Factors for Non–AIDS-Defining Cancers Keri N. Althoff 1 ; Stephen J. Gange 1 ; Chad Achenbach 2 ; Lisa P. Jacobson 1 ; Angel M. Mayor 3 ; Michael J. Silverberg 4 ; Amy Justice 5 ; Richard Moore 6 ;Yuezhou Jing 1 ; Kelly Gebo 6 On behalf of the North American AIDS Cohort Collaboration on Research and Design

1 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, US; 2 Northwestern University, Feinberg School of Medicine, Chicago, IL, US; 3 Universidad Central del Caribe, Bayamon, US; 4 Kaiser Permanente Northern California, Oakland, CA, US; 5 Veterans Affairs Connecticut Healthcare System and Yale Schools of Medicine and Public Health, New Haven, CT, US; 6 Johns Hopkins University School of Medicine, Baltimore, MD, US Background: The increased burden of non-AIDS-defining cancer (NADC) in HIV-infected adults is likely driven by both HIV-related and other cancer risk factors. The objective of this study is to estimate the population attributable fraction (PAF) for smoking and HIV-related risk factors for NADC, interpreted as the proportion of NADC that could be avoided in HIV-infected adults if all participants had the reference group exposure level. Methods: Adults ( ≥ 18 years) participating in one of 16 contributing cohorts to the North American AIDS Cohort Collaboration on Research and Design who were observed for validated NADC diagnosis from January 1, 2000 to December 31, 2009 were included in this analysis. HIV-related risk factors included CD4 count <200 cells/mm 3 , HIV RNA ≥ 200 copies/mL, and clinical AIDS diagnosis. Hepatitis B (HBV) and C (HCV) infections and smoking were also examined. Data on alcohol use, BMI, and HPV infections were not currently available. Risk factors were measured at study entry, with the exception of time-dependent CD4 count and HIV RNA. Cox proportional hazard models with piecewise constant baseline hazard functions were used to estimate adjusted hazard ratios (aHR) and 95% confidence intervals ([,]). The PAFs for the modifiable risk factors of interest were estimated using the methodology described by Laaksonen, et al . Results: Among 39,554 adults who contributed 159,914 person-years, there were 592 incident cancer outcomes distributed as 101 (17%) lung, 96 (16%) anal, 60 (10%) prostate, 54 (9%) Hodgkin, 42 (7%) liver, and 42 (7%) breast cancers. No other cancer type represented more than 5% of the NADC. At baseline, participants who developed NADC were older and had greater proportions with a history of smoking, dyslipidemia, HBV, HCV, and an AIDS diagnosis compared to those without NADC. The PAFs for the variables in the final model can be seen in Figure 1. After excluding lung cancers from the analysis, the PAF for smoking was 39% [23%, 52%].

146

CROI 2015