CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

763 LONG-TERM OUTCOMES OF AN INTEGRATED MATERNAL AND CHILD HIV CARE TRIAL IN SOUTH AFRICA Tamsin K. Phillips 1 , Kirsty Brittain 1 , Yolanda Gomba 1 , Pheposadi Mogoba 1 , Allison Zerbe 2 , Elaine J. Abrams 2 , Landon Myer 1 1 University of Cape Town, Cape Town, South Africa, 2 ICAP at Columbia University, New York, NY, USA Background: The MCH-ART trial demonstrated that co-located maternal HIV and routine paediatric care, integrated in maternal and child health (MCH) services through weaning, increased retention in care and viral suppression (VS) through 12m postpartum. Long-term outcomes after leaving integrated services are not known. Methods: To assess long-term outcomes in the MCH-ART cohort, an additional study visit was conducted at 36-60m postpartum including interviews and viral load (VL) testing. Provincial electronic health records were accessed to ascertain deaths and retention regardless of attending the additional visit. The primary outcomes were: 1) retention (any ART visit, pharmacy dispensing, CD4 count or VL test) and 2) VS (<50 copies/mL), in the 12m prior to the study visit or prior to the median time postpartum at the study visit for women who did not attend. Predictors of retention and VS were assessed using Poisson regression reported as adjusted risk ratios (aRR) with 95% confidence intervals (CI). Results: Of 471 women enrolled in MCH-ART (Jun 2013-Dec 2014), 450 (96%) were followed in routine medical records (11 withdrew fromMCH-ART or refused further follow-up; 10 deaths were ascertained) and 353 (75%) completed the additional study visit (May 2017-Apr 2018; median 44m postpartum). Of 450 women followed, 63%were retained in HIV care; among 368 women with either study or routine VL available, 56% had VS. The MCH-ART intervention effect observed at 12m postpartum did not persist at 24m and 36m postpartum; loss from care appeared similar in both trial arms after the last visit in the integrated clinic (Figure). The risk of non-retention was increased by nadir CD4 cell count >350 cells/µL (vs ≤200; aRR 1.43 95% CI 0.97-2.12) and late gestation at presentation for antenatal care in MCH-ART (>20 vs ≤20 weeks; aRR 1.36 95% CI 1.05-1.76). The risk of VL≥50 was increased by late gestation at presentation (aRR 1.27 95% CI 1.00-1.62) and increasing log VL at delivery (aRR 1.25 95% CI 1.13-1.38); findings persisted in sensitivity analyses assuming all missing VLs were VS or not. Conclusion: This long-term follow-up of the MCH-ART trial suggests that the benefits of integrated postpartumMCH care attenuate rapidly after postpartum transfer and are lost by 36-60m postpartum. The substantial non-retention and loss of VS observed in this cohort is concerning and interventions to support women on ART beyond pregnancy and breastfeeding are urgently needed.

Poster Abstracts

764 EVALUATION OF GUIDELINES FOR VL MONITORING IN PREGNANCY & BREASTFEEDING: A SIMULATION Maia Lesosky 1 , Janet M. Raboud 2 , Tracy Glass 1 , Sean S. Brummel 3 , Andrea L. Ciaranello 3 , Judith S. Currier 4 , Shaffiq Essajee 5 , Diane V. Havlir 6 , Catherine A. Koss 6 , Anthony Ogwu 7 , Roger L. Shapiro 3 , Elaine J. Abrams 8 , Landon Myer 1 1 University of Cape Town, Cape Town, South Africa, 2 University Health Network, Toronto, ON, Canada, 3 Harvard Medical School, Boston, MA, USA, 4 University of California Los Angeles, Los Angeles, CA, USA, 5 UNICEF, New York, NY, USA, 6 University of California San Francisco, San Francisco, CA, USA, 7 Botswana Harvard AIDS Institute Partnership, Gabarone, Botswana, 8 Columbia University, New York, NY, USA Background: There are global concerns about adherence to ART in pregnant and breastfeeding (P&BF) women and subsequent elevated viral load (eVL) and MTCT risk. Intensified VL monitoring for P&BF women has been proposed in guideline recommendations but not evaluated systematically. Methods: We developed a stochastic individual patient simulation of VL in 10000 P&BF women, modelled weekly from conception through 2y postpartum. Parameters were from published evidence and model outputs were calibrated against data from studies of ART in P&BF (PROMISE, PROMOTE, MmaBana and MCHART). Simulation settings were that 50% of women initiated ART in pregnancy (median 22w gestation (IQR, 16-28)) and 50%were on ART prior to conception (86%<50 c/mL at 1stANC) with modelled ART adherence. Delivery was at median 38w (IQR, 37-40); breastfeeding was for median 40w (IQR, 29-49). We applied to the same simulated population 5 different guidelines for VL monitoring, including adaptations for P&BF women when stated. Guidelines were compared on coverage of VL testing in P&BF, the proportion of eVL in the simulation successfully detected, and the cumulative VL experienced by the time of detection. Results: Coverage of VL monitoring in P&BF varied widely by guidelines (Table). By 24m postpartum, 92% of women initiating ART achieved VL<50 c/mL, and 18% of these subsequently experienced transient or extended eVL >1000 c/ mL. Specific recommendations for testing at either a fixed gestation (WHO) or a fixed period after initiation (PHS) achieved >95% testing in pregnancy; other guidelines led to 59-83% antenatal testing; and with no special stipulation only 16% of women received an antenatal test under Malawian guidelines. Guidelines calling for monitoring in BF (SA, Kenya) had >70% testing during BF compared to 30-40% among guidelines that did not (WHO, Malawi). Only a small proportion of simulated episodes of eVL>1000 c/mL were successfully detected by monitoring (range, 11-29%); guidelines with more frequent testing in P&BF led to shorter delays from the onset of eVL to detection as well as lower cumulative VL before detection. In sensitivity analyses, findings were robust to realistic variations in the simulated population. Conclusion: Without guidance specific to P&BF women, <1 in 5 women would receive antenatal or postnatal VL monitoring. However even with specific guidance, current guidelines yield suboptimal detection of eVL. Further research

CROI 2019 295