CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

respondent’s wife had been screened (aOR 0.97, 95% CI: 0.77, 1.22) (Table 1). Men with more progressive gender views about sexual behaviors (higher GEM scores) were more likely to have a partner who had been screened (aOR 1.46, 95% CI: 1.00, 2.13) (Table 1). Conclusion: Men in this study recognized the high burden and threat of cervical cancer. However, important gaps in knowledge and a strong role in decision- making may limit access to potentially life-saving services for their wives. Our findings suggest that cancer control programs should engage male partners, given their critical role in women’s decisions about use of cervical cancer prevention.

626 IMPACT OF UNIVERSAL ART ACCESS ON KAPOSI SARCOMA: RESULTS FROM THE ICONA COHORT Alessandra Bandera 1 , Alessandro Cozzi-Lepri 2 , Antonella Cingolani 3 , Francesca Ceccherini Silberstein 4 , Giulia Marchetti 5 , Stefano Rusconi 6 , Valentina Svicher 4 , Roberto Cauda 3 , Francesco Castelli 7 , Andrea Gori 1 , Andrea Antinori 8 , Antonella D'Arminio Monforte 5 , for the Icona Foundation Study Group 1 University of Milan, Milan, Italy, 2 MRC Clinical Trials Unit at UCL, London, UK, 3 Catholic University of the Sacred Heart, Rome, Italy, 4 University of Rome Tor Vergata, Rome, Italy, 5 Azienda Ospedaliera San Paolo, Milan, Italy, 6 Luigi Sacco University Hospital, Milan, Italy, 7 ASST Spedali Civili di Brescia, Brescia, Italy, 8 IRCCS Lazzaro Spallanzani, Rome, Italy Background: The widespread introduction of effective ART reduced the burden of AIDS-related Kaposi Sarcoma (KS), even if KS does still occur also in individuals with well-controlled HIV infection. Methods: We included naïve HIV-infected individuals (PLHIV) enrolled in the ICONA cohort over 1997-2019. Prevalent cases were PLHIV with a diagnosis of KS prior and up to 30 days after enrolment. Incident cases were defined as new KS diagnoses occurring after ART initiation. Patients’ characteristics at the date of enrolment were compared by prevalent KS status and associations identified by logistic regression modelling. In the subset of people KS-free at enrolment, standard Kaplan-Meier curves were used to model time from ART initiation to development of KS and a Cox regression model to identify factors associated with this outcome. A similar analysis was performed in PLHIV with prevalent KS to identify factors associated with their risk of KS relapse or death after ART (clinical failure). Results: Among 17,742 PLHIV enrolled in the ICONA cohort over 1997-2019, 248 prevalent KS cases and 36 incident KS cases were identified. Prevalent cases were mostly male (93%), with median age of 45 years (IQR 37-53) and median CD4+ count of 76/mmc (IQR 24-193). No significant differences in prevalence (by year of enrolment) and incidence (by year of ART initiation) of KS were observed (Figure). At multivariable logistic regression, the only factor independently associated with prevalent KS was mode of HIV transmission (MSM versus PWID, adjusted odds ratio [aOR]: 5.24 (1.35, 20,39)). In contrast, factors independently associated with the risk of incident KS were pre-ART CD4+ count (adjusted relative hazard (aHR): 0.57 (0.42, 0.77) for 100 cells/mmc higher) and mode of HIV transmission (MSM versus HS, aHR: 3.82 (1.62, 9.02)). Over 1,316 PYFU, clinical failure after ART introduction was observed in 52 prevalent cases (29 deaths and 23 relapses) with an incidence rate of 3.9% (95% CI: 2.9-5.1). However, none of the considered factors showed an association with the risk of clinical failure. Conclusion: Despite universal ART access, we did not observe a reduction of KS prevalence and incidence in recent years. The strong association of pre-ART CD4+ count with incidence of KS in ART-treated PLHIV strengthens the role of immune competence in KS. Further KSHV and HIV immune-virological characterization is warranted to better identify factors associated with KS occurrence in PLHIV.

625 LUNG CANCER INCIDENCE AND RISK FACTORS DIFFER BY HISTOLOGY AMONG HIV+/– VETERANS Elizabeth Chiao 1 , Jennifer R. Kramer 1 , Yongquon Dong 1 , Peter A. Richardson 1 , Sarah T. Ahmed 1 , Donna L. White 1 , Christine Hartman 1 , Jose M. Garcia 2 , Suchismita Raychaudhury 1 , Eric A. Engels 3 1 Baylor College of Medicine, Houston, TX, USA, 2 University of Washington, Seattle, WA, USA, 3 National Cancer Institute, Bethesda, MD, USA Background: People living with HIV (PWH) have high risk for developing lung cancer (LC) and poor treatment outcomes. Predominant molecular alterations and treatment algorithms differ for each of the lung cancer histologic types, which include small cell lung cancer (SCLC), and non-small cell lung cancer (NSCLC), and 2 common NSCLC histologic subtypes squamous cell (SC), and adenocarcinoma (AC). Few studies have evaluated epidemiologic differences by histology in PWH. Methods: In a retrospective cohort study, Veterans diagnosed with HIV between 10/1/99 and 12/31/2016 were identified and matched to a 2 to 1 gender, age and year of care HIV-negative cohort. Both HIV+/- veterans were followed from index date until LC diagnosis, death or 12/31/2016. LC cases and histologic types were identified using the VA Corporate Data Warehouse and medical record review of individuals with LC ICD-9/10 codes. We identified 46604 HIV+ and 88783 HIV- veterans who met eligibility criteria. We calculated cumulative LC incidence rates by histologic types and used Joinpoint software for modeling trends. Cox regression analyses were used to identify risk factors for specific LC histologic types and subtypes among PWH. Models were adjusted for age, race, gender, year of index HIV, smoking, baseline CD4 count, and percent undetectable HIV viral load. Results: A total of 931 incident cases of LC were ascertained among HIV+ and 1206 among HIV-. The overall incidence rate (IR) of SCLC was 20.43/100,000 among HIV+ veterans and 21.37/100,00 among HIV-, and the incidence rate ratio (IRR) was 0.96 (0.84 – 1.09). Among the NSCLC subtypes, the IRs for AC was the highest for both HIV+ and HIV- (93.52/100,000 vs 49.82/100,000, IRR was 1.88 CI: 1.75 – 2.01), and the IRs for SC were lower for both HIV+ and HIV- (67.7/100,000 vs 38.3/100,000, with an IRR of 1.77 CI: 1.63 – 1.92). Fig. 1 shows the joinpoint analysis of IRs per 3-year intervals for AC and SC for HIV+/- veterans. In multivariable analysis of PWH by LC histology, we found that baseline CD4 count >200 was not significantly protective for AC (HR 1.05 CI: 0.67 – 1.63, p=0.83) and was marginally protective for SC (HR 0.61 CI:0.35 – 1.05, p=0.08). Conclusion: The IRs of SC and AC NSCLCs but not SCLC are higher among PWH. The IRs of AC and SC have remained stable over time for both HIV+/- veterans.

Poster Abstracts

CROI 2020 226