Temporal Trends of Population Viral Suppression in the Context of Universal Test and Treat: The ANRS 12249 TasP Trial in Rural South Africa

Joseph Larmarange, Mamadou Diallo, Nuala McGrath, Collins Iwuji, Mélanie Plazy, Rodolphe Thiébaut, Frank Tanser, Till Bärnighausen, Joanna Orne‐Gliemann, Deenan Pillay, François DabisSSRN 2019 · DOI

TL;DR

This study investigates the impact of the "Universal Test and Treat" (UTT) strategy on population viral suppression (PVS) in rural South Africa. While UTT significantly increased PVS in both intervention and control arms, the gains were similar, suggesting that the strategy's effectiveness is heavily influenced by factors beyond immediate ART initiation, such as robust testing and linkage to care.

What Problem Does This Paper Address?

The paper addresses a critical question in HIV/AIDS public health: how effectively does the "Universal Test and Treat" (UTT) strategy, which aims to maximize population viral suppression (PVS) by initiating antiretroviral therapy (ART) for everyone living with HIV (PLHIV) regardless of their CD4 count, translate into real-world population-level impact? Specifically, the ANRS 12249 TasP trial in rural South Africa observed a lack of significant difference in cumulative HIV incidence between UTT intervention and control arms, despite the intervention's focus on early ART. This study aims to disentangle the temporal trends of PVS within the trial, exploring whether the observed PVS changes can explain the null effect on HIV incidence and identifying which components of the UTT strategy contribute most to PVS.

Methods

The study employed a cluster-randomized trial design with two arms: an intervention arm and a control arm. Both arms received six-monthly home-based HIV counseling and testing (RHBCT). The key difference lay in ART initiation: the intervention arm offered immediate ART regardless of CD4 count, while the control arm followed national guidelines (initially CD4 ≤350, later ≤500).

To assess PVS, researchers meticulously tracked residency, HIV status, and HIV care status for all participants daily. PVS was defined as the proportion of all resident PLHIV (aged ≥16) who were on ART and virally suppressed. They used mixed linear models to analyze temporal trends in PVS, adjusting for sociodemographic changes at the cluster level. Sensitivity analyses were conducted to assess the impact of missing HIV status data.

Datasets

The primary dataset was the ANRS 12249 TasP trial database, which included:

  • Trial registrations and exits
  • Uptake and results of home-based rapid HIV testing
  • ELISA HIV serological results from dried blood spots (DBS)
  • Clinic visits, ART prescriptions, and viral loads of PLHIV seen in trial clinics

Additional data were sourced from:

  • National Health Laboratory Service (NHLS) for viral loads and CD4 counts from local government clinics.
  • AHRI clinical database (ACCDB) for ART clinic visits and prescriptions from local government clinics.

Data from these sources were matched using probabilistic scores based on identifying information.

Key Insights and Ideas

  • Beyond ART Initiation: The study highlights that while immediate ART initiation is a core component of UTT, its direct impact on PVS might be less significant than comprehensive strategies for maximizing testing and linkage to care.
  • The Power of RHBCT and Linkage: The significant increase in PVS, particularly in the intervention arm, was strongly associated with the implementation of repeat home-based HIV testing and local trial clinics, suggesting these efforts were crucial in identifying and connecting individuals to care.
  • Context Matters: Calendar time played a role, with changes in national treatment guidelines (e.g., the increase in CD4 threshold for ART initiation) influencing PVS, especially in the control arm.
  • The 90-90-90 Challenge: The findings underscore that achieving the ambitious 90-90-90 targets (90% diagnosed, 90% on ART, 90% virally suppressed) requires addressing operational and implementational challenges that limit access to care, which are often context-specific.

Key Results

  • Significant PVS Increase: PVS increased substantially in both arms over the trial period. In the intervention arm, it rose from 23.5% to 46.2% (+22.8%), and in the control arm, from 26.0% to 44.6% (+18.6%).
  • Similar PVS Gains: Despite the intervention, the increase in PVS was statistically similar between the two arms (p=0.514). This suggests that the additional benefits of immediate ART initiation were modest compared to the gains achieved through enhanced testing and linkage strategies implemented in both arms.
  • Drivers of PVS Increase: The primary drivers of PVS increase were "time since cluster opening" (measuring the impact of RHBCT and local clinics) and, to a lesser extent, calendar time.
  • Limited Effect of Trial Arm: The direct effect of the trial arm (i.e., immediate ART initiation) on PVS increase was positive but limited.

Limitations and Future Work

  • Data Limitations: The study acknowledges limitations in precisely estimating care status and ART status for individuals followed in local governmental clinics, as proxy indicators were used. Matching algorithms for combining data from different sources also have inherent limitations.
  • Lower Bounds for PVS: Due to data limitations, the estimated proportions of PLHIV being in care and virally suppressed are considered lower bounds.
  • Power for Incidence: The trial was not powered to estimate temporal trends of HIV incidence, which limits the ability to definitively link PVS changes to incidence reductions.
  • Future Directions: The authors suggest that future research should focus on addressing operational and implementational challenges that hinder access to care and treatment, as these are critical for achieving population-level HIV control. They also highlight the need to understand how population dynamics (migration, new infections) influence the HIV care cascade.

Summary updated

BibTeX
@misc{larmarange2019temporal,
  author = {Joseph Larmarange and Mamadou Diallo and Nuala McGrath and Collins Iwuji and Mélanie Plazy and Rodolphe Thiébaut and Frank Tanser and Till Bärnighausen and Joanna Orne‐Gliemann and Deenan Pillay and François Dabis},
  title = {{Temporal Trends of Population Viral Suppression in the Context of Universal Test and Treat: The ANRS 12249 TasP Trial in Rural South Africa}},
  year = {2019},
  doi = {10.2139/ssrn.3324749},
  howpublished = {SSRN},
}
Figure