HIV treatment outcomes among people with initiation CD4 counts >500 cells/µL after implementation of Treat All in South African public clinics: a retrospective cohort study

INTRODUCTION: The World Health Organisation recommends to Treat All people with HIV, irrespective of CD4 count. However, people with CD4 counts >500 cells/µL may be asymptomatic and therefore less motivated to adhere to antiretroviral therapy (ART). We aimed to assess whether people initiated with CD4 counts >500 cells/µL had worse treatment outcomes compared to those initiated at lower CD4 counts. METHODS: We performed a retrospective cohort study among non‐pregnant adults initiating ART at eight public clinics in South Africa between September 2016, when Treat All was implemented, and August 2017. We assessed whether initiation CD4 count >500 cells/µL was associated with the outcomes of attrition (death, lost to follow‐up or treatment interruption >180 days), and viraemia >1000 copies/mL, by twelve months using Cox proportional hazards and Poisson regression models. RESULTS AND DISCUSSION: Among 4952 patients initiating ART, the median age was 32.4 years (interquartile range (IQR) 27.2 to 39.7), 58.9% were women and 30.3% had an initiation CD4 count >500 cells/µL. After twelve months, 3382 (68.3%) were retained in care, 303 (6.1%) had transferred to another clinic, 1010 (20.4%) were lost to follow‐up, 232 (4.7%) had a treatment interruption >180 days and 25 (0.5%) were known to have died. Overall, 1267 experienced attrition at a median time of 91 days (IQR 23 to 213), with 302 of these (23.8%) experiencing attrition immediately after their ART initiation visit. Among those in care at twelve months with viral load results, 4.6% had viraemia. In multivariable analysis, the hazard of attrition was similar between patients newly eligible for ART with CD4 counts >500 cells/µL compared to those with CD4 ≤500 cells/µL (adjusted hazard ratio 1.03, 95% confidence interval (CI) 0.90 to 1.17). The risk of viraemia was lower among patients with CD4 counts >500 cells/µL compared to those with CD4 ≤500 cells/µL (adjusted risk ratio 0.58, 95% CI 0.37 to 0.92). CONCLUSIONS: After implementation of Treat All in South African public clinics, we found that patients newly eligible for ART with initiation CD4 counts >500 cells/µL had comparable or better outcomes compared to those with lower CD4 counts. These finding support ongoing implementation of Treat All in our setting.