CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

year increase 0.98, 95%CI 0.97-0.99), HIV status disclosure (OR 0.31, 95%CI 0.11-0.83) and ART adherence skills score (OR 0.70 per 10% increase, 95%CI 0.53-0.93); we observed a trend for depression (OR 2.14, 95% CI 0.91-5.02). Marital status, self-reported adherence, history of side effects, social support, history of intimate partner violence or distance to clinic were not associated with virologic failure. Conclusion: Virologic failure during pregnancy varied widely by MCH clinic, raising concerns about disparities in programmatic implementation of Option B+ in Kenya. Additional support for women who are young, have not disclosed, report difficulty with adherence behavioral skills, and/or with depression may be important to optimize Option B+ efficacy. 814 ART ADHERENCE AMONG PREGNANT AND NON-PREGNANT WOMEN IN SOUTH AFRICA AND UGANDA Lynn T. Matthews 1 , Catherine Orrell 2 , Bosco M. Bwana 3 , Stephen Asiimwe 4 , Gideon Amanyire 3 , Nicholas Musinguzi 3 , Anna Cross 2 , Kathleen Bell 1 , Alexander C. Tsai 1 , Christina Psaros 1 , David R. Bangsberg 5 , Jessica E. Haberer 1 1 Massachusetts General Hospital, Boston, MA, USA, 2 University of Cape Town, Cape Town, South Africa, 3 Mbarara University of Science and Technology, Mbarara, Uganda, 4 Kabwohe Clinical Research Center, Kabwohe, Uganda, 5 Oregon Health and Sciences University, Portland, OR, USA Background: With the advent of Option B+, women living with HIV (WLWH) are offered more streamlined care, but data suggest ongoing challenges to ART adherence and retention in care. We conducted research to understand patterns and correlates of ART adherence among WLWH accessing public sector clinics offering Option B+ as standard of care in Uganda (UG) and South Africa (SA). Methods: We prospectively observed pregnant and non-pregnant women initiating ART with CD4>350 cells/ml during 2015-2017 in southwestern, UG and Cape Town, SA. Women were seen at 0, 6, and 12 months for socio-behavioral questionnaires and HIV RNA levels. Adherence to ART was monitored in real- time (Wisepill). Those lost to follow-up were considered non-adherent/virally unsuppressed; data were censored at death/disenrollment. Predictors of ART adherence were assessed by multivariable linear regression in site-specific models. Results: Among 439 enrolled women, 205 pregnant and 234 non-pregnant women had median age 26 (IQR 23-29) vs. 31 (IQR 24-41), CD4 554 (IQR 430-683) vs. 436 (IQR 396-472), with 54 (27%) vs. 68 (36%) reporting a partner living with HIV, and 85 (43%) vs. 96 (51%) reporting serostatus disclosure to partner, respectively. In UG, median adherence to ART for pregnant and non-pregnant women was similar (90% [IQR 75-98] vs. 89% [IQR 73-96], p=0.17), but lower among pregnant women in SA (36% [IQR 15-70] vs. 76% [IQR 41-91], p<0.001). For women in UG, adherence was independently associated with older age (b=1.1 [0.30, 1.9]), employment (b=14.5 [5.3, 23.8]), depression (b=26.2 [14.3, 38.1], and alcohol use (b=-15.5 [-29.2, -1.7]). Substance use (b=-19.8 [-35.2, -4.4]) was the only independent predictor of adherence among non-pregnant women in SA. For pregnant women in SA, adherence was independently associated with marriage (b=24.0 [0.4, 47.5]), depression (b=-31.5 [-53.4, -9.5]), gestational age (b=-1.2 [-2.4, -0.1] per week), and non-disclosure to partner (b=-8.4 [-28,11]). In UG, 96 (90%) and 100 (86%) pregnant and non-pregnant women were virally suppressed at 12 months. In SA, viral suppression was achieved by 36 (58%) of pregnant and 91 (91%) of non-pregnant women. Conclusion: Effective methods to promote mental health and support within partnerships, while mitigating alcohol and substance use for WLWH accessing care during pregnancy are crucial for the health of women, their children, and their families, particularly in urban South Africa.

Poster Abstracts

815 EARLY CLINICAL EVENTS AFTER ART INITIATION IN PREGNANCY INFLUENCE VIRAL LOAD OUTCOMES

Jasantha Odayar 1 , Elaine J. Abrams 2 , Allison Zerbe 2 , Tamsin K. Phillips 1 , Landon Myer 1 1 University of Cape Town, Cape Town, South Africa, 2 ICAP at Columbia University, New York, NY, USA Background: There is growing concern around adherence and virologic outcomes following antiretroviral therapy (ART) initiation in pregnancy. Incident clinical events after ART initiation, including side effects, new symptoms and ‘minor’ complaints of pregnancy, may influence short-term ART outcomes, but data are few. Methods: Consecutive women initiating TDF+FTC+EFV in routine antenatal care in Cape Town, South Africa were followed with viral load (VL) monitoring through 12 months postpartum. Data on clinical events during the antenatal period were abstracted from routine records. For each woman, the first episode of a sign, symptom or diagnosis (including side effects, pregnancy disorders and other illnesses) was included. Poisson methods were used to assess the association between clinical events and VL at delivery and through 12 months postpartum. Results: In 553 women enrolled (median age 28 years; median CD4 count 349 cells/µL; median pre-ART VL 4.0 log[10] copies/ml; median gestation 20 weeks; 258 person-years observation; 1819 clinic visits), 48% (n=263) had at least one clinical event (23%, 14% and 11% had 1, 2 and ≥3 events, respectively). There were 512 clinical events recorded in the cohort, with peak incidence in the first 8 weeks after ART initiation (Figure). Clinical events were significantly more likely at CD4 counts <100 and 100-200 than >200 cells/µL (IRR 1.94, 95% CI 1.21–3.13; and 1.38, 95% CI 1.04–1.83, respectively). Overall, 18% of clinical events were dermatologic, 15% neurological, 14% gastrointestinal, 14% genital, and 12% upper respiratory tract. At delivery (median ART duration 120 days) 72%women had viral suppression (VS) (VL<50 copies/mL). Adjusting for age, enrolment VL, and ART duration, delivery VS was marginally less likely in women experiencing a clinical event antenatally (IRR 0.97, 95% CI 0.90–1.04). This association was amplified in women on ART ≥16 weeks at delivery (IRR 0.93, 95% CI 0.86–0.99) and women who experienced ≥3 clinical events before delivery (IRR 0.93, 95% CI 0.78–1.10). Only 7% of all clinical events resulted in referral to a higher level of care; this was not associated with VS at delivery. There were no associations between clinical events in pregnancy and virologic outcomes in the postpartum period. Conclusion: Incident clinical events after ART initiation occur commonly during pregnancy and appear associated with viral suppression in the short term; these may warrant specific attention in patient counselling and support interventions.

CROI 2018 307