CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

treat analysis (Table), earlier ART resulted in a 47% hazard reduction for any cancer (p=0.002), 42% reduction for virus-unrelated NADC (p=0.032) and 60% reduction for ADC (p=0.007). No reduction was observed for the smallest cancer group of virus-related NADC (p=0.91). Per-protocol results were similar, although somewhat attenuated (Table). Conclusion: Results from this large North American HIV cohort suggest that the burden of certain cancer groups may be substantially reduced as more HIV providers and patients adopt ART guidelines for immediate initiation.

Poster and Themed Discussion Abstracts

599 TIMING OF CART INITIATION INFLUENCES CANCER RISK AMONG WOMEN LIVING WITH HIV Kate Salters 1 , Monica Ye 2 , Neora Pick 3 , Angela Kaida 1 , SamWiseman 4 , Shahab Jabbari 2 , Mark Hull 2 , Jean A. Shoveller 2 , Julio S. Montaner 2 , Robert S. Hogg 2 1 Simon Fraser Univ, Burnaby, British Columbia, Canada, 2 BC Cntr for Excellence in HIV/AIDS, Vancouver, British Columbia, Canada, 3 Oak Tree Clinic, Vancouver, British Columbia, Canada, 4 St. Paul’s Hosp, Vancouver, British Columbia, Canada Background: Since the advent of combination antiretroviral therapy (cART), the incidence rate of AIDS-defining malignancies (ADM) in people living with HIV has declined, largely due to improved immune function. However, these trends are not gender-stratified and may obscure changes in cancer risk specific to women living with HIV (WLWH), in particular ADM such as cervical cancer, and non-ADM. We assessed the impact of early cART initiation on the incidence of all-type cancer, ADMs, and non-ADM among WLWH. Methods: The Comparison of Outcomes and Service Utilization Trends (COAST) study is comprised of a retrospective, administrative database of people living with HIV, both on and off cART, in British Columbia, Canada between January 1996 and March 2013. Our analytic study population was restricted to women with HIV diagnosis and incident cancer cases after HIV diagnosis were identified by ICD-O codes. We conducted a Poisson regression to determine correlates of all-type cancer, ADM, and non-ADM diagnosis, and calculated the cancer incidence rate (per 1,000 person-years (PYs)) by CD4 cell count at cART initiation (≤200, 200-349 and ≥350 cells/mm3). We also calculated the attributable fraction of risk associated with CD4 cell count at cART initiation. Results: Among 1,660 WLWH, 50 women were diagnosed with cancer after HIV diagnosis (31 ADM and 19 non-ADM). Earlier initiation of cART (≥350 cells/mm3) was associated with lower all-type cancer incidence (Relative Risk (RR): 0.33 [95% CI: 0.16, 0.70]) and non-ADM diagnosis (RR: 0.15 [95%CI: 0.03, 0.64]), but not ADM diagnosis, compared to cART initiation at CD4 of ≤200 cells/mm3. After adjusting for age at HIV diagnosis, the incidence of all-type cancer and non-ADMwas 5.55 [95% CI: 3.89, 7.91] cases per 1,000 PY and 2.50 [95% CI: 1.47, 4.27] cases per 1,000 PY, respectively for those with CD4 cell count of ≤200 cells/mm3 compared to those with baseline CD4 of 350 cells/mm3. The attributable fraction of risk for all-type cancer incidence was 63.6% and 82.06% for non-ADM incidence for those with CD4 cell count of ≤200 cells/mm3 at cART initiation. Conclusion: Early initiation of cART was protective against all-type cancer and non-ADM cancer among WLWH, however no additional benefit was observed for reducing incidence of ADM cancers. In the context of ‘Treatment as Prevention’, this study suggests there are significant oncological health benefits of early treatment initiation for WLWH. 600 CANCER RISK AMONG HIV-INFECTED PEOPLE IN THE US DURING THE MODERN TREATMENT ERA Raul Ulises Hernandez-Ramirez 1 , Meredith S. Shiels 2 , Robert Dubrow 1 , Eric A. Engels 2 1 Yale Univ, New Haven, CT, USA, 2 NCI, Rockville, MD, USA Background: HIV-infected people have an elevated risk for many cancers, mainly AIDS-defining cancers (ADC) and virus-related non-ADC (VRNADC), but not for most virus- unrelated non-ADC (VUNADC). Although ADC risk declined substantially after the introduction of effective antiretroviral therapy (ART) in 1996, trends for other cancers have been less clear. We describe recent patterns of cancer risk among a cohort of 448,258 HIV-infected people from 9 US states (1996-2012). Methods: We assembled this cohort using the HIV/AIDS Cancer Match Study, a population-based linkage of HIV and cancer registries. We calculated standardized incidence ratios (SIRs) to measure risk relative to the general population. For cancers with SIRs that were elevated and changing over time (based on unadjusted Poisson models), we further assessed calendar trends using piecewise models that adjusted for demographic characteristics. Results: Between 1996-1999 and 2009-2012, the proportion of follow-up time among HIV-infected people 50+ years-old increased from 12% to 29%. Overall cancer risk was elevated (SIR 1.69, 95%CI 1.67-1.72; N=21,294 cases). SIRs were elevated (p<0.001) for ADC (14.0) and VRNADC (5.4) but not for VUNADC as a group (0.9). SIRs were elevated for each individual ADC (Kaposi sarcoma [KS, 498], non-Hodgkin lymphoma [NHL, 11.5], and cervix [3.2]), most VRNADC (oropharynx/tonsil [1.6], anus [19.1], vagina [3.6], vulva [9.4], penis [5.3], and liver [3.2]), and 10 other cancers, e.g., lung [2.0], larynx [2.1], scrotum [6.8], conjunctiva [5.6], and myelodysplastic syndrome [2.0]. Risk was not elevated for other common cancers, e.g., colorectum, breast, and prostate. SIRs for ADC, VRNADC, VUNADC, and some individual cancers decreased significantly across calendar periods. Among 10 cancers selected for detailed assessment, 6 cancers (KS, diffuse large B-cell lymphoma, central nervous system NHL, anus, liver, and lung) showed significantly decreasing SIRs across 1996-2012 (Fig. 1).

CROI 2017 253