Abstract:
Background: Globally, approximately 10-15% of newly diagnosed people with HIV (PWH)
have drug-resistant viruses. In low-and middle-income countries (LMICs), the rise of HIV-1 drug
resistance (HIVDR) and virological non-suppression pose a major challenge to end the HIV
epidemic. Several studies and the World Health Organization (WHO) have outlined an increase
in pretreatment drug resistance (PDR) in LMICs since the global antiretroviral treatment (ART)
scale-up. High levels of PDR in such circumstances endanger the HIV response by lowering viral
suppression rates, increasing the number of AIDS-related deaths per year, and increasing HIV
incidence. Moreover, the effects of baseline laboratory markers including HIV-1 viral load (VL)
and PDR on treatment outcomes have been rarely studied since the introduction of dolutegravir
(DTG)-based regimens as a first-line ART in Ethiopia.
Objective: To characterize pretreatment HIV-1 drug resistance and determine virological and
treatment outcomes among newly diagnosed people with HIV in eastern Ethiopia between
October 2020 and July 2022.
Methods: At baseline, a total of 235 newly diagnosed PWH who started on first-line ART were
enrolled into the cohort from 15 health facilities, and then, followed up for six months. A
questionnaire and checklist were used to collect demographic and clinical data. Five milliliters of
venous blood was collected at baseline and six months after ART initiation. The Abbott m2000sp
automated sample-preparation system was used to extract HIV RNA. The baseline and six-month
VL were also quantified using the Abbott Real-Time HIV-1 assay. The samples were amplified
with PCR using an in-house assay. The amplicons were sequenced with the Sanger dideoxy
method. Drug resistance mutation (DRM) profiles were analyzed and interpreted using the
calibrated population resistance (CPR) and Stanford HIVDR database. A recombinant
identification program and REGA subtyping tool were used to infer genetic diversity. PhyML
version 3.0 was used to generate a maximum likelihood phylogenetic tree, which was then
visualized using the iTOL online tool. STATA version 14.0 was used to perform logistic
regression analysis. In the final model of analysis, statistical significance was defined as a p value of <0.05.
Results: The median age was 35 years (IQR: 6-62) and the majority, 70.6% were females. All
samples with baseline VL (≥1000 copies/mL) were eligible for HIV-1 pol gene sequence, of
which the genotyping success rate was 78.4%. Among the isolates successfully sequenced, three
HIV-1 strains were detected, of which 97.1% had HIV-1 subtype C, 1.4% A1C, and 1.4% CF1
recombinant. According to the Stanford HIVDR algorithm, 21.7% of people whose HIV isolates
xvii
were genotyped had at least one drug associated PDR mutation, whereas CPR reported 14.5%
DRMs. Non-nucleoside reverse transcriptase inhibitors (NNRTIs) had the highest PDR mutation
rate (13.0%), followed by nucleoside reverse transcriptase inhibitors (NRTIs) (7.2%) and
protease inhibitors (PIs) (2.9%). The most commonly observed major DRMs were: NNRTIs
(K103N and G190A), NRTIs (D67G and L210W), and PIs (L90M and I54S). The prevalence of
mutations conferring intermediate or high-level resistance to NNRTIs and NRTIs was 10.1%.
Being bedridden at baseline had a higher likelihood of harboring PDR mutations (AOR=5.4;
95% CI: 1.53-30.7) than those who were in a working functional status. The incidence of
acquired drug resistance (ADR) among viremic PWH and samples sequenced at six months was
80% (4/5) and the K103N mutation was detected twice. Among the PWH enrolled at baseline,
68.5% were still in ART care, 16.6% loss to follow-up (LTFU), 8.9% transferred out (TO), and
5.9% died at six months. Of the 161 retained on ART, 8.7% (95% CI: 5.2-14.2) failed to achieve
viral suppression. The baseline predictors of unsuppressed VL were age ≤30 years (AOR= 8.9,
95% CI: 1.85, 42.8), history of substance use (AOR=7.5, 95% CI: 1.4, 39.9), and VL greater than
4-log10 copies/mL (AOR= 12.6, 95% CI: 1.65, 96.5). Baseline WHO clinical stage I (AOR
=3.93, 95% CI: 1.34–11.6), low VL (AOR =3.67, 95% CI: 1.09–12.4), and higher body weight
(AOR=1.04, 95% CI: 1.01–1.07) were predictors of LTFU, but non-functional status (AOR
=10.02, 95% CI: 1.9–51.3) was the single predictor of death.
Conclusions: There is high accumulation of PDR mutations, predominantly to NNRTIs. In this
cohort, virological non-suppression was found to be optimal but still slightly lagged behind the
third 95% target. The six months LTFU and death rates among patients are relatively high,
accounting for nearly a quarter of all attrition. Therefore, further studies on drug resistance
surveillance are needed to evaluate the long-term effects of prolonged buildup of resistance on
the current ART regimens, and to design appropriate interventions to curb the HIV epidemic.
Greater efforts are required to implement baseline VL testing in order to improve treatment
outcomes and foster the attainment of the 3rd 95% of the UNAIDS–target. Moreover, for the
baseline factors significantly associated with PDR, unsuppressed VL, LFTU, and death, targeted
interventions and broader research are required to investigate the causes of the underlying
factors, mitigate poor treatment outcomes, and improve the health of the patients in the ART
program in the study area.