CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

285 CD4/CD8 RATIO AS A PREDICTOR OF HIV-ASSOCIATED CANCERS IN CNICS Chad J. Achenbach 1 , Brian Joyce 1 , Lifang Hou 1 , Elizabeth Hibler 1 , Jeffrey N. Martin 2 , W. C. Mathews 3 , Richard D. Moore 4 , Thibaut Davy-Mendez 5 , Benigno Rodriguez 6 , Kenneth H. Mayer 7 , Michael Saag 8 , Mari Kitahata 9 , for the CFAR Network of Integrated Clinical Systems (CNICS) 1 Northwestern University, Chicago, IL, USA, 2 University of California San Francisco, San Francisco, CA, USA, 3 University of California San Diego, San Diego, CA, USA, 4 Johns Hopkins University, Baltimore, MD, USA, 5 University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, 6 Case Western Reserve University, Cleveland, OH, USA, 7 Fenway Health, Boston, MA, USA, 8 University of Alabama at Birmingham, Birmingham, AL, USA, 9 University of Washington, Seattle, WA, USA Background: CD4/CD8 ratio is available in routine HIV practice and has been associated with immunoscenece, aging and cancer. We aimed to evaluate associations between low CD4/CD8 ratio, prior to and on ART, and individual cancer types. Methods: We studied persons with HIV (PWH) in care 1996-2014 at 8 CNICS sites across the US, who initiated ART and had a pre-ART CD4/CD8 ratio value and at least 6 months of follow up. We assessed quartiles of the lowest CD4/ CD8 ratio prior to and the highest CD4/CD8 ratio on ART as predictors of cancer overall and individual types from ART initiation to cancer event or last clinic visit using Cox regression models adjusted for age, sex, race, time on ART, HCV co-infection, tobacco and alcohol use. Results: Among 10,817 PWH with 63,514 person-years on ART, 80%were male, 51% non-white race, 61%MSM, 17% IDU and 21% had HCV co-infection. Prior to ART, median nadir CD4 count was 229 cells/µl (IQR 89, 361), max CD8 count was 980 cells/µl (IQR 680, 1380), lowest CD4/CD8 ratio 0.29 (IQR 0.14, 0.47) and HIV RNA 4.9 log 10 copies/mL (IQR 4.3, 5.4). On ART, 91% achieved HIV RNA suppression <200 copies/mL and median highest CD4/CD8 ratio achieved was 0.76 (IQR 0.47, 1.11). 529 PWH developed invasive cancer: 93 NHL, 86 KS, 53 lung, 43 anal, 38 prostate, 33 Hodgkin lymphoma, 20 liver, 19 breast, 18 colorectal, 16 oropharynx, 11 melanoma and 99 others. After adjustment, pre-ART CD4/CD8 ratio <0.14 (lowest quartile) was significantly associated with greater risk of cancer overall (HR 1.8; 95%CI 1.4-2.3), anal cancer (HR 4.7; 95%CI 1.5-13.0) and NHL (HR 3.9; 95%CI 1.7-8.8) compared to ratio ≥0.47 (highest quartile). On ART, CD4/CD8 ratio <0.47 (lowest quartile) was also associated with KS, lung cancer and Hodgkin lymphoma compared to ratio ≥1.11 (highest quartile)(Table). We did not find statistically significant associations between CD4/CD8 ratio prior to or on ART and melanoma, colorectal, breast, kidney, prostate or liver cancer. Conclusion: As has been observed with other age-related diseases, CD4/CD8 ratio is a potential biomarker routinely obtained in HIV care that could be used for prediction of certain cancers and risk stratification for HIV-associated cancer screening strategies. 286 CAUSES OF DEATH AFTER CANCER DIAGNOSIS AMONG PLHIV ON ART: COHORT COLLABORATION Adam Trickey 1 , Michael John Gill 2 , Dominique Costagliola 3 , Sophie Grabar 3 , Amy C. Justice 4 , Suzanne Ingle 1 , Julia Del Amo 5 , Antonella D’Arminio Monforte 6 , Fabrice Bonnet 7 , Joerg Janne Vehreschild 8 , Ferdinand Wit 9 , Katharina Grabmeier- Pfistershammer 10 , Leah Shepherd 11 , Ramon Teira 12 , Jonathan Sterne 1 1 University of Bristol, Bristol, UK, 2 University of Calgary, Calgary, AB, Canada, 3 INSERM, Paris, France, 4 Yale University, New Haven, CT, USA, 5 Institute of Health Carlos III, Madrid, Spain, 6 University of Milan, Milan, Italy, 7 University of Bordeaux, Bordeaux, France, 8 Cologne University Hospital, Cologne, Germany, 9 Stichting HIV Monitoring, Amsterdam, Netherlands, 10 Innsbruck Medical University, Innsbrusk, Austria, 11 University College London, London, UK, 12 Hospital Sierrallana, Torrelavega, Spain Background: People living with HIV (PLHIV) are more likely to develop AIDS- defining malignancies (ADMs) and several non-ADMs (NADMs) than the general population. However, there is a lack of information on cause-specific mortality after diagnosis of cancer among PLHIV. Methods: We investigated causes of death within 5 years of first cancer diagnosis in PLHIV enrolled in 10 European and North American HIV cohorts that are part of the Antiretroviral Cohort Collaboration (ART-CC). Eligible adults

were aged ≥16 years, started ART between 1996-2015 and were subsequently diagnosed with cancer. We used CoDe classifications of cause of death https:// chip.dk/Tools-Standards/CoDe/About. We calculated cause-specific mortality rates (MR) per 1000 years following diagnosis of specific cancers and compared all-cause MR during 2006-15 with 1996-2005, for ADMs and NADMs. Results: After 4209 cancer diagnoses (ADM=2162, NADM=2047) among 84167 PLHIV, there were 1451 deaths within 5 years. Of 604 PLHIV who died after diagnosis of ADM, 292(48%) deaths were due to an ADM while 467/847 (55%) deaths after diagnosis of NADMwere due to an NADM. MR were higher for diagnoses between 1996-2005 compared with 2006-15: ADMs 102 (95% CI 92-113) vs 88 (78-100) and NADMs 213 (191-239) vs 184 (169-200). The table shows mortality rates for the 7 most commonly diagnosed cancers: these were higher after diagnoses of NADMs than ADMs and were very high for lung, liver, non-Hodgkin’s lymphoma, and head and neck cancers. Patterns of cause- specific mortality suggest that cause of death was likely to have been from the diagnosed cancer for head and neck and lung cancer. A substantial proportion of deaths from liver cancer had been classified as due to viral hepatitis by our process for assigning CoDe cause of death classifications. Conclusion: Among ART-treated PLHIV diagnosed with cancer, mortality rates and causes of death varied according to the type of cancer, with highest mortality for the NADMS liver cancer and lung cancer. Among those who died within 5 years of a diagnosis of an NADM there was a high chance that death was from cancer and not from AIDS.

Poster Abstracts

287 AIDS PROGRESSION AND NON-AIDS DEATH IN PEOPLE WITH HIV FOLLOWING CANCER TREATMENT

Keri Calkins 1 , Geetanjali Chander 2 , Corinne Joshu 1 , Kala Visvanathan 1 , Anthony T. Fojo 2 , Catherine R. Lesko 1 , Richard D. Moore 2 , Bryan Lau 1 1 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA, 2 Johns Hopkins University School of Medicine, Baltimore, MD, USA Background: CD4 and HIV viral load trajectories following cancer treatment in people with HIV (PWH) may influence both AIDS and cancer progression. We estimate the association between longitudinal CD4 and viral load and the competing events of incident AIDS-defining illness/AIDS-death (ADI) and non-AIDS death (NAD) after cancer diagnosis in PWH. Methods: We analyzed data from 204 PWH in the Johns Hopkins HIV Clinical Cohort diagnosed with an incident cancer from 1997-2014. Initial cancer treatment was categorized into chemotherapy/radiation and surgery without chemotherapy/radiation based on their observed immunosuppressive effects. We used shared parameter joint longitudinal survival models to estimate the cause-specific hazard ratios (csHR) and subdistribution hazard ratios (sdHR) for the associations between longitudinal CD4 and viral load with ADI and NAD. Models adjusted for confounders, including SEER estimates of 5-year mortality based on cancer type and stage and ART use. Results: PWH were on average 50 years old at cancer treatment (IQR=43- 55); 59% received chemotherapy/radiation. When compared to surgery, chemotherapy/radiation was associated with a rapid initial decline of 149 CD4 cells/μL (95% CI=87,211) among patients initiating treatment with CD4≥350, and 64 cells/μL for those with CD4<350. A 100-cell increase in longitudinal CD4 resulted in a csHR of 0.84 (95% CI 0.60-1.13) and an sdHR of 0.91 (95% CI= 0.69-1.18) for ADI, adjusted for longitudinal viral load. Although not statistically significant, this implies that higher CD4 was protective against ADI. A 100-cell increase in longitudinal CD4 was associated with a lower hazard of NAD (csHR=0.77, 95%CI= 0.64-0.91; sdHR=0.73, 95%CI= 0.61-0.85), adjusted for longitudinal viral load. Increased longitudinal viral load after cancer treatment in PWH was not significantly associated with an increased hazard of ADI or NAD when accounting for changes in longitudinal CD4 after cancer treatment.

CROI 2019 104