The prognostic value of HIV viral load in predicting viraemic outcomes of post-partum women on antiretroviral therapy: a secondary analysis of two randomised controlled trials from Gugulethu, Cape Town

Egan, Daniel

The prognostic value of HIV viral load in predicting viraemic outcomes of post-partum women on antiretroviral therapy: a secondary analysis of two randomised controlled trials from Gugulethu, Cape Town

Thesis / Dissertation

2023

Abstract

Background: Vertical transmission of HIV continues to be a significant health concern for HIV-infected women in pregnancy and postpartum, especially in Southern Africa, despite decreasing rates of infection. While better access to effective and acceptable antiretroviral therapy (ART) regimens has helped to improve control of viraemia in this group, ongoing vertical transmission continues despite being entirely preventable. The dynamics of transient viraemia, even in previously well-controlled individuals on ART, are not fully understood and may contribute to ongoing mother-to-child transmission (MTCT) despite therapy that is otherwise optimal. Methodology: Two randomised controlled trials (RCTs), the MCH-ART and PACART trials, were conducted in an antenatal clinic in Cape Town, South Africa, between 2013-2016 and 2016-2019 to investigate improvements to postnatal HIV care through use of integrated postnatal maternal and child clinics and adherence clubs, respectively. The present study conducted a pooled secondary analysis of data from these two trials to investigate the dynamics of HIV viraemia in these women and to estimate the prognostic strength of viral load (VL) measurements during the postpartum period. Sensitivity, specificity, predictive values, likelihood ratios and odds ratios for VL measurements were calculated at different cut-offs to predict later VL >1000 copies/ml. Cox proportional hazards models were used to estimate the risk of two consecutive measurements of VL >1000 copies/ml over the study period, based on starting viral load. Results: 883 HIV-infected pregnant women were followed-up for a median of 22.1 months (IQR: 17.7- 24.1), with a longer median length of follow-up in the PACART cohort (24.0 months) compared to the MCH-ART group (17.9 months). 826 (93.5%) of participants recorded at least one episode of VL <50 copies/ml, with 694 (78.6%) demonstrating VL <50 at the first study visit postpartum. 306 women (37%) demonstrated at least one episode of viraemia in the range of 50-999, while 307 (37.2%) experienced one or more episodes of VL >1000. As a predictor of future VL >1000, measurable viraemia >50 copies/ml was a specific (87.5%) but not sensitive (68%) test, with a clinical odds ratio of 14.8 (95% CI: 12.4-17.8). Similarly, current VL >200 was a good predictor of VL >1000 at the next visit, with an odds ratio of 26.1 (95% CI: 21.4-31.8), a sensitivity of 64%, and specificity of 94%. Cox proportional hazards analysis found that viraemia in the range of 50-999 copies/ml at the first study visit post-partum was associated with an increased hazard of two consecutive VL >1000 over the study period, when compared to participants with VL <50 at the same timepoint (HR: 1.75; 95% CI: 1.3-2.4; p<0.001). Conclusion: Episodes of viraemia in HIV-infected postpartum women remain common despite improving ART regimens and public HIV services. These represent windows of increased risk during which vertical transmission is more likely to occur. VL testing remains a useful tool in monitoring viraemia and may be helpful for identifying those who are more likely to develop substantial and prolonged viraemic episodes. These findings highlight one of the remaining problems perpetuating vertical transmission of HIV: unmeasured, untreated viraemia.

Keywords

Public Health and Family Medicine

Reference:

Collections

Masters

Full item page