CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

632 HIV Testing, Status, and Treatment Among Patients at the Uganda Cancer Institute

Matine Ghadrshenas 1 ; Rachel A. Bender Ignacio 1 ; Daniel H. Low 1 ;WarrenT. Phipps 2 ; Jackson Orem 3 ; Ann Duerr 2 ; Corey Casper 2 1 Univ of Washington, Seattle, WA, USA; 2 Fred Hutchinson Cancer Rsr Cntr, Seattle, WA, USA; 3 Uganda Cancer Inst, Kampala, Uganda

Background: HIV increases the incidence and mortality of cancer; knowledge of HIV status and treatment is essential for the management of patients with HIV-associated malignancies (HIVAM). In Uganda, where the prevalence of HIV infection is >7%, the incidence of AIDS-defining cancers (ADCs) -including Kaposi sarcoma, Non-Hodgkin lymphoma, and cervical cancer– is high, and Non-AIDS defining cancers (NADCs) are increasingly common among HIV-infected (HIV+) persons. We determined how often cancer providers documented the HIV status and clinical parameters of HIV infection among patients presenting for care at the Uganda Cancer Institute (UCI), the primary cancer treatment facility for a catchment area of 100 million. Methods: Medical records of patients aged ≥18 who registered at the UCI between May-August 2015 were abstracted for demographics, cancer and HIV parameters. We calculated binomial proportions and used χ 2 tests and logistic regression to evaluate factors associated with HIV testing, HIV-positivity, and antiretroviral therapy (ART) usage. Results: Among 556 patients, 30% had a potential ADC. 67.8% of charts documented HIV status. Of those with documented HIV status, 137 (36%) were HIV+, and 58% of HIV+ individuals had an ADC. The documented HIV prevalence in NADCs was 24%. Men were 1.75-fold more likely to be HIV positive (95% CI 1.15-2.68, p=0.009), however, women were more likely to have undocumented HIV status (RR 1.32, p=0.009). Women accounted for 54.6% of all patients; 36% of women lacked HIV test results, including 40% of women with cervical cancer. HIV+ patients were younger compared with those known to be HIV-negative (median age 41 vs 50, p<0.001). Among those with documented HIV infection, 62% had a CD4 count recorded in the chart. The median CD4 count among persons with ADCs was 300 cells/ml (interquartile range 114-395) compared with NADC (median 353, IQR 185- 601), p=0.08. There was no difference in the proportion of HIV patients receiving ART prior to UCI registration between those with ADC vs. NADC (86%), p=0.45. Conclusions: HIV prevalence was 5 times higher in Ugandan cancer patients with documented status than the general population. Though the majority of cancer patients had HIV testing performed, gaps remained in documenting HIV status, CD4 count and ART usage, even among patients with ADCs. This study highlights opportunities to educate cancer clinicians in Africa on the burden of HIV in cancer patients and the importance of managing both diseases in patients with HIVAM.

Poster Abstracts

633

Low CD4/CD8 Ratio As a Predictor of Cancer Risk in HIV-Infected Persons Keith M. Sigel 1 ; Brinda Emu 2 ; Lesley Park 3 ; Kristina Crothers 4 ; Cynthia Gibert 5 ; Matthew Goetz 6 ; Sheldon Brown 7 ; Maria Rodriguez-Barradas 8 ; JanetTate 9 ; Robert Dubrow 10 1 Icahn Sch of Med at Mount Sinai, New York, NY, USA; 2 Yale Univ Sch of Med, New Haven, CT, USA; 3 Stanford Univ, Stanford, CA, USA; 4 Univ of Washington Sch of Med, Seattle, WA, USA; 5 VA Med Cntr, Washington, DC, USA; 6 VA Greater Los Angeles Hlthcare System, Los Angeles, CA, USA; 7 James J. Peters VA Med Cntr, Bronx, NY, USA; 8 Michael E. DeBakey VA Med Cntr and Baylor Coll of Med, Houston, TX, USA; 9 VA Med Cntr, West Haven, CT, USA; 10 Yale Univ Sch of PH, New Haven, CT, USA Background: Low CD4/CD8 ratios in HIV-infected (HIV+) persons may represent dysfunctional immune activation and have been found to be associated with increased non-AIDS associated mortality. Using data from a large cohort of HIV+ Veterans, we investigated the relationship between cumulative CD4/CD8 ratio and risk of incident non-AIDS defining (NADC) and AIDS defining (ADC) cancers. Methods: We linked the Veterans Aging Cohort Study (VACS) to the Veterans Affairs Central Cancer Registry to yield a cohort of 26,115 HIV+ subjects followed for a minimum of 2 years, during 1997-2012. We categorized the primary exposure of interest, longitudinal CD4/CD8 ratio, using previously identified cut points for non-AIDS event risk (<0.7 vs ≥0.7). We calculated observation time from VACS enrolment to the earliest of: pathologically confirmed incident cancer, loss to follow-up, or death. Cancers were classified as NADC or ADC and then further subclassified by viral association and anatomic site. Our CD4/CD8 ratio measure was the 18-month simple moving average (SMA), lagged by 6 months (to minimize reverse causality). These lagged SMAs were then evaluated as time-updated covariates in separate Cox proportional hazard regression models for each cancer type (ADC, NADC, virus-related NADC, and the most common constituents of these groups). Models were adjusted for demographics, smoking, drug and alcohol use disorders, recent CD4 count, and hepatitis C virus infection. Results: We identified 1,259 and 229 incident NADCs and ADCs, respectively, in our cohort. Baseline median CD4/CD8 ratio and CD4 count was lower for subjects who were eventually diagnosed with both NADCs and ADCs compared to patients who never developed cancer. In Cox regression models (Table 1) cumulative CD4/CD8 ratio <0.7 was significantly associated with increased risk of NADCs (hazard ratio [HR]: 1.2; 95% CI: 1.1-1.4) but not with ADCs or the subset of NADCs specifically associated with viral co-infections after adjustment for potential confounders. Among NADCs only lung cancer (HR: 1.5; 95% CI: 1.1-2.1) and anal cancer (HR: 2.0; 95% CI: 1.1-3.8) were associated with low cumulative CD4/CD8 ratios after adjustment; there was no association with other NADCs. Conclusions: In our large, antiretroviral therapy-era HIV cohort, we found that cumulative exposure to low CD4/CD8 ratio was associated with lung cancer and anal cancer risk after adjustment for potential confounders including recent CD4 count.

257

CROI 2016