CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

990 Increased Levels of HIV Baseline Drug Resistance Testing & Early Treatment Initiation Hong-Ha M. Truong 1 ; Sharon Pipkin 2 ; Robert Grant 1 ;Teri Liegler 1 ; Susan Scheer 2 1 Univ of California San Francisco, San Francisco, CA, USA; 2 San Francisco Dept of PH, San Francisco, CA, USA

Background: National HIV treatment guidelines have evolved greatly in the past decade. Current guidelines recommend early initiation of antiretroviral treatment (eiART) and baseline drug resistance testing (bDRT) upon diagnosis and linkage to care. bDRT and eiART patterns were assessed with respect to national treatment guidelines. Methods: San Francisco residents diagnosed with HIV/non-AIDS between 2001 and 2012 who were linked to care at publicly-funded facilities were included in the analysis (N=2,565). Cases were stratified by era of treatment guidelines based on diagnosis year: Era 1=2001-03 (n=657), Era 2=2004-06 (n=716), Era 3=2007-09 (n=640) and Era 4=2010-12 (n=552). bDRT and eiART were defined as occurring within 3 months of diagnosis. Results: The proportion of cases with bDRT increased from 2001 to 2012 (p<0.001). The nadir was 4% in 2001, increased to 9% in 2002, dipped to 8% in 2003, rose to 12% in 2004, and then continued to increase to a peak of 56% in 2012. bDRT increased across the 4 eras, from 7% to 15% to 33% to 49% (p<0.001). The proportion of cases with eiART increased from 2001 to in 2012 (p<0.001). The proportion was 16% in 2001, decreased over the next 3 years to a nadir of 4% in 2004, and then began to increase steadily to a peak of 45% in 2012. eiART increased across the 4 eras, from 11% to 6% to 16% to 36% (p<0.001). eiART was associated with an increased likelihood of bDRT (p<0.001). Gender, age, race/ ethnicity and HIV transmission risk were not associated with bDRT overall; persons <20 years and Latinos were more likely to have been tested in Era 4 (p<0.05). Conclusions: Baseline drug resistance testing increased steadily from 2004 to 2012, though not all genotyped cases initiated treatment within 3 months of diagnosis. Nevertheless, baseline genotyping can be useful for informing regimen selection upon the subsequent decision to initiate ART, since a transmitted resistant mutant present early in the course of infection may still be in viral archives but no longer detectable as the most abundant variant. eiART increased substantially in Era 4 which may reflect the adoption of treatment as prevention, whereas the observed decrease in late Era 1 and in Era 2 may stem from concerns during that time about the toxicities associated with ART. Lack of disparities in baseline drug resistance testing by demographic or risk group may be due to federal and municipal support that facilitated access to primary care services. 991 HIV Prevalence and Care in the New York State Department of Corrections Lou Smith 1 ; Brenda L. Moncur 2 ; Bridget Anderson 1 ; JamesTesoriero 3 ; Carl J. Koenigsmann 4 ; Heidi Reukauf 3 ; Deepa Rajulu 1 ; Daniel A. O’Connell 1 1 New York State Dept of Hlth, Albany, NY, USA; 2 Hlth Rsr, Inc, Albany, NY, USA; 3 New York State Dept of Hlth, AIDS Inst, Albany, NY, USA; 4 Dept of Corrections and Community Supervision, Albany, NY, USA Background: In 1988, an unlinked seroprevalence study showed that nearly 1 in 5 incoming inmates to New York State (NYS) Department of Corrections and Community Supervision (DOCCS) facilities had HIV infection. Enormous challenges have been addressed in DOCCS facilities with extensive programs for HIV testing, medical care, prison medical staff education, transition planning and stigma reduction. Continuing studies have shown decreasing HIV seroprevalence in incoming inmates with 2.4% of males and 3.7% of females seropositive in 2012. With changing NYS laws on use of surveillance data and NYS’s “Ending the Epidemic” initiative, this project was undertaken to understand the extent and experience of diagnosed HIV infected persons under custody and to further intervention opportunities. Methods: An electronic file of DOCCS inmates under custody inclusive of HIV medical diagnosis codes was matched to the NYS Department of Health (DOH) HIV Registry using probabilistic matching software. CD4 lymphocyte (CD4) and viral load (VL) data routinely reported to the HIV registry were used as a proxy for HIV-related medical care. Evidence of HIV care was defined as >1 CD4 or VL. Prevalence was based on inmates under custody on 12/31/2014, with HIV care assessed for July – December, 2014. Results: 49,224 persons were under custody on 12/31/2014. 1,109 matched to cases in the HIV Registry, for an HIV prevalence of 2.3% (men 2.2%, women 3.7%). HIV care was assessed for 887 who were diagnosed before July 2014 and under custody continuously from July to December, 2014. Medical diagnoses codes documented that 796 (89.7%) of those were known to DOCCS as having HIV. 776 (87.4%) had evidence of HIV care, and 774 (87.2%) had a VL. VL closest to yearend was <200 copies/ml in 721 (81.2%). Conclusions: This work represents the first direct measure of the prevalence of diagnosed HIV in those under custody in NYS DOCCS. Among those with diagnosed HIV, nearly 9 of 10 had disclosed their diagnosis to DOCCS and were receiving care. Viral suppression of those receiving any HIV care (90.6%) compares favorably to a similar statewide rate of 82% in 2013. The match provides important opportunities for further public health interventions, including interventions with out-of-care persons who have previously disclosed status to DOCCS, intervention with diagnosed HIV-infected incarcerated persons not in care and not known to DOCCS, and NYSDOH follow-up to assess and intervene for linkage to care post-incarceration. 992 Frequency and Duration of Churn Among Persons LivingWith HIV inWashington, DC Maggie Dorr 1 ; Amanda Castel 2 ; Kerri Dorsey 2 ; Adam Allston 3 1 George Washington Univ, Washington, DC, USA; 2 George Washington Univ Milken Inst Sch of PH, Washington, DC, USA; 3 District of Columbia Dept of Hlth, HIV/AIDS, Hepatitis, STD, and TB Administration, Washington, DC, USA Background: Measuring retention in care is an essential component of the HIV care continuum. However, traditional retention measures often require extensive resources but also tend to focus on one point in time, consequentially failing to adequately capture the care dynamics of those who cycle in and out of care. It is unclear whether this cycling, or “churn,” results in poorer clinical outcomes. The objective of this analysis was to characterize the frequency of churn and duration of gaps in care among persons living with HIV, as well as identify individual characteristics associated with varying churn patterns. Methods: Analyses were restricted to routinely collected HIV and laboratory surveillance data for persons diagnosed with HIV in the District of Columbia prior to 2008 and in care at the end of 2013 with ≥1 visit in 2008 and ≥1 visit in 2013. Churn was defined as being out of care for ≥6 months. Frequencies, univariate, bivariate, and nonparametric analyses, as well as multivariate logistic regression were conducted to assess associations between individual characteristics and churn. Results: Among 3,364 people, 91% (n=3,064) had evidence of churn. The mean duration of churn was inversely related to the number of gaps: persons with one gap in care had

Poster Abstracts

a mean gap of 63 weeks compared to 36 weeks among persons with seven gaps in care. Among the 3,064 individuals who churned, 26% had one gap, 28% two gaps, 22% three gaps, and 25% had 4-7 gaps in care between 2008 and 2013. Frequency of churn varied significantly by sex (p=0.04), race/ethnicity (p<0.01), age (p<0.01), and duration of HIV infection (p=0.01). Multivariate logistic regression found that, compared to persons with no churn, those 45-64 years old were significantly less likely to experience two or more gaps in care than those under 35 years of age. Blacks were also significantly more likely to churn than whites. Persons diagnosed with HIV ≤1 year prior to their 2008 visit were significantly more likely to churn compared to those diagnosed more than 15 years prior. There was a significant change between reported CD4 and viral loads between 2008 to 2013 across all churn categories. Conclusions: These results add to our understanding of the impact of taking periodic breaks from care and help quantify the large amount of churn observed among people living with HIV. Our data on churn duration may also assist in more efficient allocation of relinkage resources by focusing on those persons who are younger and have been more recently infected.

422

CROI 2016