Trends in HIV-1 pretreatment drug resistance and HIV-1 variant dynamics among antiretroviral therapy-naive Ethiopians from 2003 to 2018: a pooled sequence analysis

Although decades have passed since the start of the pandemic, the Human Immunodeficiency Virus (HIV) remains a major public health threat globally. The COVID-19 pandemic exacerbated the challenge of HIV infection due to overstretched healthcare delivery, competition for resources (including financial and healthcare-providing professionals), and the synergistic effect of co-infection [1]. According to the UNAIDS 2022 report, approximately 38.4 million people were living with HIV (PLWH) worldwide during the year 2021. Among these, the eastern and southern African regions bear the highest burden of the illness, covering 54% (20.6 million) of the global estimate. Similarly, being situated in this region, Ethiopia is among the most highly affected countries, with 12,000 AIDS-related deaths and 12,000 new HIV infections in the same year [2].

Anti-retroviral therapy (ART) has been scaled up globally from 7.8 million in 2010 to 28.7 million people in December 2021, resulting in a reduction of the burden associated with the pandemic [2]. While this is an outstanding progress in terms of scale-up, the development of pretreatment HIV-1 drug resistance (PDR), which comes in parallel with ART, is becoming a challenge for ART effectiveness [3]. ART scale-up in Sub-Saharan Africa, where drug resistance testing is not routinely available, is also similarly associated with an increasing prevalence of PDR [4, 5].

In Ethiopia, the publicly funded ART program was started in 2005, and the number of people using ART has now scaled up from 58,405 in the beginning to around 436,000 in 2017 [6‐9]. Despite this, however, recent reports show that the PDR is increasing [10]. Therefore, epidemiologic data on PDR is crucial to helping establish optimal empiric ART regimens, especially in countries like Ethiopia, where treatment options are limited and HIV drug resistance testing is absent. To the best of our knowledge, no review study shows the trend of PDR among ART-naive HIV-positive individuals in Ethiopia. Hence, this review has examined the trend of PDR among ART-naïve HIV-positive individuals along with the HIV-1 variant dynamics over the years in Ethiopia.

We first identified 823 research articles and excluded 788 of the studies based on title and abstract review. The rest 25 full-text articles were further reviewed based on the eligibility criteria, which led to the selection of 12 articles for analysis, three of which were removed due to the unavailability of sequence data (Fig. 1).

Currently, PDR has become a challenge to the success of ART, which makes HIV prevention very difficult. According to the WHO's recent report, PDR prevalence is increasing globally [12]. The same is true for low and middle-income countries, which have shown a substantial increment recently [4]. Therefore, determining PDR at the population level is crucial for optimizing patient and population-level treatment outcomes that will consequently help achieve the third WHO/UNAIDS 90–90–90 target. This review provided the first nationally representative description of the PDR trend along with the HIV variant dynamics over the years.

The pooled prevalence of PDR among ART-naïve individuals in Ethiopia in this systematic review was 4.8%, which is low as per the WHO HIV drug resistance classification [11]. This finding is in line with a similar study from China [13, 14], reports from Tanzania [15], and Kenya [16]. It is, however, lower than the rate of PDR reported from South Africa [17], Cameroon [18], Namibia [19], Brazil [20], Italy [21], and England [22].

The NNRTI-associated PDRMs were the most frequently observed mutations in this review, which corroborates observations from different studies across the globe [23‐28]. This finding supports the recent transition of the ART regimen in Ethiopia from NNRTI to WHO-recommended integrase strand transfer inhibitor-based regimens. The K103N mutation, which is a non-polymorphic mutation that confers a high-level resistance against efavirenz (EFV) and nevirapine (NVP) [29, 30], was found in a high proportion (17.6%; 9/51) from the NNRTI associated PDRMs. This is similar to reports from Malawi [26], Israel [23], Brazil [20], Tanzania [15], Namibia [19], and Italy [21]. The G190A (13.7%; 7/51) mutation, which is equally selected in persons receiving NVP and EFV [31, 32] was also a common NNRTI-associated mutation, followed by the K101E mutation (7.8%; 4/51), Y188C (7.8%; 4/51) and G190S (5.9%; 3/51) (Fig. 2). The K101E and Y188C mutations are associated with reduced susceptibility to NVP, Etravirine (ETR), and Rilpivirine (RPV) [33‐35], while Y188C and G190S reduce susceptibility to NVP and EFV [33, 35‐37].

Among the NRTI-associated mutations, K219Q and L210W mutations which are selected by the thymidine analogs zidovudine (AZT) and Stavudine (d4T), were reported in high frequencies (9.8% (5/51) and 7.8% (4/51)), respectively. In addition to these mutations, others like T215S (5.9%; 3/51) and T215FIS (3.9%; 2/51) were also observed in this review, although in relatively lower frequencies. T215S/I mutation is associated with an increased risk of virologic failure associated with AZT or d4T usage, while T215F confers reduced susceptibility to all currently approved NRTIs except emtricitabine (FTC) and lamivudine (3TC) [38]. It is known that AZT and d4T are commonly used in the country similar to other resource-limited settings; hence, the presence of the mentioned mutations underscores the need for routine genotypic resistance testing.

Protease inhibitor (PI) mutations are usually observed at lower frequencies in different studies [23‐28, 39]. Likewise, the frequency of PI mutations is low (0.8%) in this review. The mutations observed are I85V (5.9%; 3/51), which was the most prevalent PI mutation, followed by M46I (3.9%; 2/51) and F53L (3.9%; 2/51). The I85V is a non-polymorphic mutation selected by Nelfinavir (NFV) and Atazanavir (ATV) [40, 41] while the M46I and F53L are associated with reduced susceptibility to ATV and Lopinavir (LPV) [42, 43]. M46I, a non-active site mutation in HIV-1 protease, has been clinically associated with saquinavir (SQV) resistance in HIV patients [44]. Altogether, the mentioned mutation profile in this review indicates that regimens containing NFV, ATV, and LPV are more likely to be effective relative to their NNRTI and/or NRTI counterparts.

Concerning the PDR prevalence per year, the overall PDR prevalence was observed to increase from 3.3% in 2003 to 9.8% in 2018 (Fig. 3). The accumulation of the PDRMs over time might have led to the upward trend of PDR over the years. In line with this finding, a similar increment has also been reported in other studies from Mozambique [45] and South Africa [17]. In addition, the NNRTI-associated PDRMs also increased over time (from 2.19% in 2003 to 7.84% in 2018) (Fig. 4), which is once again similar to observations from other studies around the world [23‐28, 39].

The current study confirmed the HIV-1 subtype C virus as the predominant clade in Ethiopia, similar to other eastern and southern African countries like South Africa [46], Mozambique [45], Malawi [26], and Botswana [47]. This class accounts for 97% of the pandemic in the country. Despite being low, the influx of other variants, including HIV-1 subtype C-like (2%, 20/1070), recombinant of C and A1 (0.38%, 4/1070), HIV-1 subtype A (A1) (0.38%, 4/1070), HIV-1 subtype D (0.09%, 1/1070), HIV-1 CRF 02_AG (0.09%, 1/1070), and recombinant of D and A1 (0.09%, 1/1070), has been observed (Fig. 5). The occurrence of these other variants indicates the possible introduction of other HIV-1 subtypes from neighboring countries. Therefore, preventive measures should always be there to prevent the possibility of further introduction of new subtypes. The 50 transmission clusters found in the sequences analyzed suggest that many HIV-positive individuals engage in risky sexual behavior, emphasizing the necessity of a thorough understanding of the transmission dynamics of HIV-1 subtype C in the population. An increased public health intervention program aimed at these people is also necessary to stop such potentially risky behaviors that will negatively affect HIV transmission. The current review has some limitations. Firstly, all data used in the analysis were retrieved from the literature, and hence we were unable to analyze the correlation of drug resistance with demographic and clinical characteristics, as this information was not available for the majority of sequences. Secondly, in the current review, the studies were mainly from the capital city and from the northern part of the country, and hence some regions might have been underrepresented (due to absence of study), which might lower the representativeness of the study.

In this systematic review, we presented temporal trends of HIV-1 PDR among ART-naïve individuals along with HIV-1 variant dynamics in Ethiopia. Despite the decade-long pandemic, the HIV variant dynamics remained the same, with HIV-1 subtype C still accounting for the majority of the clade (> 97%). Our analysis showed that the overall pooled prevalence of PDR is borderline low (4.8%) 15 years after the rollout of ART in Ethiopia. However, the trend has been seen to be increasing over the years, which is concerning and indicates the need for routine surveillance of PDRMs along with consecutive preventive measures in the country. In addition, the observed increasing trend in NNRTI PDR supports the recent transition of the ART regimen from NNRTI to WHO-recommended integrase strand transfer inhibitor-based regimens.