Rates of initial virological suppression in patients with resistance to antiretroviral therapy

Abstract

Given increasing levels of pre-treatment HIV drug resistance in South Africa (SA), particularly to the non-nucleoside reverse transcriptase inhibitor (NNRTI) drug class, the efficacy of standardised NNRTI-based first-line regimens used routinely to achieve virological suppression in patients initiating first-line ART in SA remains unclear. As part of the joint HIV/TB prospective cohort surveillance study “Prospective Sentinel Surveillance of Tuberculosis and Human Immunodeficiency Virus in South Africa and Related Drug Resistance”, conducted between November 2014 and December 2017, a medical chart review was performed for participants enrolled in the study 6-months post initiation of antiretroviral therapy (ART) to monitor rates of retention-in-care and virological response to ART. Factors associated with having a viral load ≥1,000 copies/mL were explored amongst participants with or without detectable levels of pre-treatment drug resistance (PDR) at treatment initiation. Patients initiating ART at 9 high burden clinics in 6 South African provinces were consented and enrolled into the primary study. Standardised interviews and medical chart reviews were performed by surveillance officers at each site at time of enrolment. Of 1330 participants enrolled in the primary study, 6-month follow up data was available for 918 (69%) participants. Inclusion was found not to be at random; participants enrolled into the study 2016-2017 were significantly more likely to be included in to the secondary analysis. In addition, female participants, participants who were unemployed or had regular employment, participants with CD4 count >500 cells/µL at treatment initiation, or participants with prior ART exposure were more likely to be excluded from the secondary analysis. Modified Poisson regression was performed to determine factors associated with retention-in-care and having a 6-month viral load result ≥ 1,000 copies/mL. Sensitivity analyses were performed to assess the robustness of the findings. Of the 918 participants included in the secondary analysis, 502 (55%) participants were retained in care, whereas 416 (45%) participants had no evidence of a 6-month visit. Of these, 375 (41%) were documented as or assumed to be LTFU, 36 (4%) were transferred to a different facility, and 5 were deceased. At a multivariate level, retention in care was significantly associated with enrolment during 2016 (adjusted incidence rate ratio (aIRR) Page 4 of 62

1.31, 95% CI 1.07 – 1.61), a relationship that maintained significance throughout the sensitivity analyses. Of 502 participants retained in care, 358 (71%) had a viral load result available; of these, 305 (85%) had a VL result <1,000 copies/mL, and 53 (15%) had a VL result that was ≥1,000 copies/mL PDR at treatment initiation was significantly associated with 4 fold increased risk of VL≥1,000 copies/mL (IRR = 4.33, 95% CI 2.63 – 7.12). Being of older age (IRR = 0.43 (95% CI 0.24 – 0.78) for 15-24 years’ age group and IRR=0.09 (95% CI 0.01 – 0.56) for participants >50 years of age) and CD4 count >200 cells/µL (IRR=0.05, 95% CI 0.36 - 0.98) were associated with significantly reduced risk of VL≥1,000 copies/mL). Our analysis showed that high proportions of participants enrolled in the parent study were not retained-in-care at 6-month post-initiation of ART. However, as these figures may reflect incomplete record keeping at the facilities, the implementation of a unique patient identifier and electronic record keeping systems that can provide real-time monitoring of patients must remain a priority activity for the national programme. In addition, having a VL ≥ 1,000 copies/mL at 6-months post initiation of ART was significantly associated with having PDR at treatment start. As such, real-time monitoring of patients with viraemia to allow for rapid management response and treatment switch must remain a priority activity for improved patient management. The study is limited by difficulty of access to paper-based medical charts at facilities and inadequate record keeping in paper-based medical files. In addition, follow-up was not performed in all facilities due to unavailability of staff. As such, rates of retention-in-care and virological outcomes may be compromised by inadequate recording of the 6-month VL result into the patients’ files.