CROI 2025 Abstract eBook

Abstract eBook

Poster Abstracts

at least one resistance mutation at baseline was 32.9% (26/79) (2.5% NRTI, 24.1% NNRTI, and 3.8% PI). The median CD4 nadir was 352 (236-500). Retention rate at week 48 was 64.4%, with main reasons for non-retention being loss to follow-up (19.8%), relocation to another city (5.9%), and incarceration (5%). ART adherence was incomplete in 57.3% of participants, and 10.1% had interrupted ART more than 20 days during at least one of their visits. In the per-protocol analysis, by week 48, 96.9% of patients achieved a viral load < 50 copies/ml. Of these, two patients had discontinued ART by the time of their 48-week visit; however, no virological failures were observed during the study. In the multivariate analysis, acquisition of HIV through injection drug use was associated with lower retention in HIV care [ORa 6.1 (95% CI 1.75-21.3) p = 0.005]. Adverse events were reported in 68.3% of patients, though none required drug withdrawal, and 95.3% were of mild intensity. Conclusions: The feasibility of a “test, treat, and retain” strategy, combined with BIC/FTC/TAF, proved to be effective in achieving and maintaining viral suppression, despite challenges related to adherence and interruptions in ART. New models of HIV care should emphasize social support to improve retention in care for people with HIV from injection drug use. Forgiveness of Dolutegravir-Based Regimens Using Medication Possession Ratio: Andhra Pradesh, India Ramesh Allam 1 , Manjula Thogarucheeti 2 , Nalini Chava 3 , Ramesam Ganti 3 , Jayakrishna Kurada 3 , Vijay V. Yeldandi 3 , Rajendra Prasad 3 , Sreeramu S. Chakravarthy 2 , Steven Y. Hong 1 1 US Centers for Disease Control and Prevention, Delhi, India, 2 Andhra Pradesh State AIDS Control Society, Hyderabad, India, 3 SHARE INDIA, Hyderabad, India Background: Adherence to antiretroviral therapy (ART) is independently associated with viral load (VL) suppression among people living with HIV (PLHIV). Previous studies have demonstrated that dolutegravir-based regimens (DBR) have forgiveness for strict adherence to achieve VL suppression (VLS) as measured by pharmacy refill and electronic drug monitoring, where 75% and 80% adherence, respectively, can achieve VLS for 90% of PLHIV. We sought to investigate forgiveness to DBR in a programmatic setting in India using medication possession ratio (MPR). Methods: We analysed program data of PLHIV transitioned from non nucleoside reverse transcriptase inhibitors (NNRTI) or protease inhibitors (PI) to DBR or initiated on DBR (as a fixed dose combination) in Andhra Pradesh, India for at least 12 months during November 2020-September 2023 with a VL result. MPR was calculated by dividing the number of days of ART dispensed by 365 days multiplied by 100. MPR was calculated back from the date of the latest VL. We further ordinally categorized MPR (<50%, 50-79%, 80-84%, 85-89%, ≥90%). Bivariate and multivariable binomial regression analyses were used to determine factors associated with VL suppression (HIV RNA <200 copies/mL) after adjusting for all covariates and within Pearson correlations. We used Probit regression models to create dose-response plots and to estimate the adherence required to achieve VL suppression in 90% of HIV VL tests. Results: Of the 120,397 PLHIV included in the analysis, 75.8% were aged <50 years, 56.2% were women, 94.3% had VL <200 copies/mL, 58.2% had MPR ≥90% and 84.6% had been transitioned to DBR. PLHIV aged ≥50 years (adjusted odds ratio [aOR] 1.36; 95% CI: 1.28 to 1.45), female (aOR 0.91; 0.87, 0.96), and transitioned to DBR (aOR 1.17; 1.10, 1.25) had greater odds of being virally suppressed compared with persons aged <50 years, men, and initiated ART with DBR, respectively. All four categories of MPR >50% had increasing significant aOR compared to <50%. The MPR level required to achieve viral suppression in 90% of HIV viral load tests was 60% (figure). Conclusions: DBRs achieved >90% of VLS even at MPR of 60%, indicating high durability of DBRs across variable MPR levels. However, effectiveness was highest for adherence >90%.

695

Less Frequent Clinical and Viral Load Assessments During COVID-19 Did Not Increase Virologic Failure Andrew Carr 1 , Modest Nwanbila 2 , Lena Ktorza 3 , Kaitlyn McCann 3 , Goran Likevic 2 , Tara Suchak 4 , Rachael Jones 4 , Anton Pozniak 2 , Andrew Hill 5 , Marta Boffito 4 1 St Vincent's Hospital, Sydney, Australia, 2 Chelsea and Westminster Hospital, London, UK, 3 Imperial College London, London, UK, 4 Chelsea and Westminster NHS Foundation Trust, London, UK, 5 University of Liverpool, Liverpool, UK Background: Antiretroviral (ARV) guidelines (WHO, DHHS, IAS-USA, EACS, BHIVA) suggest HIV viral load (VL) monitoring every 3 to 6 months. No trial has evaluated intervals >6 months. There are few recommendations regarding the interval between patient assessments. During the COVID lockdown, clinical assessment and VL monitoring were less frequent in the Chelsea-Westminster (C-W) HIV cohort (3 London clinics, 5 regional clinics) and more clinical assessments were by phone. The primary aim of this analysis was to see if this less intensive HIV care during the COVID pandemic resulted in more virological failure (VF) than pre or post-pandemic. Methods: We compared VF rates (VL>200 cp/mL after VL<50 cp/mL) in three periods: pre-COVID (Jan 2018-March 2020, 26 months), COVID pandemic (April 2020-December 2021, 21 months) and post-pandemic (Jan 2022-October 2023, 22 months) in the C-W cohort, which prospectively records patient assessments and VLs. Survival analysis compared times to VF between periods, and Cox proportional hazards model to examine the impact of covariates (sex, age, born in UK, clinic area, ARV anchor class) on time to VF. Results: 14,897 adults (mean 47.0 yrs, 14.2% female, 39.6% UK-born, 87% HIV RNA <50 cp/mL) were included. The most common ARVs were TDF, FTC and an integrase inhibitor (INSTI). Mean intervals between VL tests were 30.8 weeks pre-COVID, 39.6 weeks during the pandemic and 27.9 weeks post-pandemic. The mean intervals between clinical assessments were 25.4, 32.5 and 23.8 weeks, respectively. Time to VF in the COVID period was similar to the pre-COVID period (OR 0.98, p>0.2) and longer than in the post-COVID period (OR 1.073, p=0.001; Figure 1). Clinical factors independently associated with VF in the pre-COVID period included CD4 count >600 (OR 0.51, p<0.001) and UK-born (OR 0.86, p=0.042), in the pandemic period were male (OR 0.66, p<0.001), age >45 (OR 2.14, p<0.001) and in the post-pandemic period were male (OR 0.61, p<0.001), age >45 (OR 1.52, p<0.001), care in London (OR 1.11, p=0.02) and face to face consultation (OR 0.15, p<0.001). For anchor class, the greatest risk of VF in all three periods was with a non-nucleoside reverse transcriptase inhibitor. Conclusions: In this population, less frequent VL monitoring and less frequent clinical assessment during the COVID pandemic were not associated with more VF. Less frequent VL monitoring and less frequent clinical assessment (provided they are face-to-face) may be possible for those with undetectable viral loads.

694

Poster Abstracts

CROI 2025 198