Interruption patterns in HIV clinic visits and treatment failure among pregnant and postpartum women in the Kabeho Study in Kigali, Rwanda
BACKGROUND: ART for HIV-positive pregnant women is recommended for life, yet most studies describe retention as a binary outcome rather than patterns of intermittent attendance รข?? a more likely description of the true nature of lifelong engagement in care and treatment. The absence of ART is strongly associated with high viral load, an indicator of treatment failure and an increased likelihood of vertical transmission of HIV. Pregnancy and postpartum are challenging periods where interruptions are common.
METHODS: The Kigali Antiretroviral and Breastfeeding Assessment for the Elimination of HIV (Kabeho) study was an observational prospective cohort of 608 HIV-positive women enrolled in their third trimester of pregnancy or within two weeks post-delivery, between April 2013 and May 2014. Interviews conducted at enrollment included questions focused on demographics, history of HIV and ART, and health and nutrition behaviors. Maternal viral load was assessed at the 24-month visit. Attendance at clinic visits were used to construct an interruption variable defined as a missed visit followed by a return to care. We examined the number and length of interruptions as predictors of treatment failure in multivariate analyses adjusting for age, months on ART, CD4 count, and ART regimen.
RESULTS: 80% of the study population had at least one interruption, and the mean number of interruptions was 1.8 per woman (SD=1.5). For women who had an interruption, the length ranged from 1-16 months with a mean of 2 months (SD=1.9). As compared to women with no interruptions, the odds of treatment failure (>1000 copies/ml) were 7 times higher among women who had an interruption of >2 months (OR=7.05, 95%CI=2.13, 23.29), and nearly 5 times higher among women with >3 interruptions throughout the study period (OR=4.95, 95%CI=1.40, 17.49). Even just one interruption (and each additional interruption) increased the odds of treatment failure by nearly 40% (OR=1.38, 95%CI=1.10, 1.73).
CONCLUSIONS: Interruptions in HIV care visits are common. Numerous and long interruptions are detrimental to viral suppression. Measuring patterns of interruption, rather than using a binary measure at one point in time, more accurately captures the fluid nature of life-time, health-seeking behavior, and can be assessed using clinic attendance data.
E. Nawar * (1), K. Andrinopoulos (2), T. Carton (3), E. Bobrow (4), A. Asiimwe (5), P. Mugwaneza (6)
(1) Elizabeth Glaser Pediatric AIDS Foundation, Washington DC, United States, (2) Tulane University School of Public Health and Tropical Medicine, New Orleans, United States, (3) Louisiana Public Health Institute, New Orleans, United States, (4) Mathematica, Washington DC, United States, (5) National Early Childhood Development Program, Kigali, Rwanda, (6) Rwanda Ministry of Health, Kigali, Rwanda