Transmitted drug resistance in ART-naive persons with acute/ early/ primary HIV infection: a systematic review and meta-analysis


 Background: In the absence of AIDS vaccine，antiretroviral therapy (ART) was the most effective tool to prevent and control the HIV pandemic. But the widespread use of ART has raised concerns about the emergence of HIV transmitted drug resistance (TDR). Acute HIV infection (AHI) was the most appropriate time to detect the spread of TDR. In this meta-analysis, our purpose was to evaluate the level of TDR in ART-naive patients with acute/ primary/ early HIV infection, and describe the critical drug-resistant mutations. Methods: We systematically reviewed 1192 studies published between January 1, 2008 and December 30, 2019 in PubMed, Web of Science, Embase, and the Cochrane Library, and selected 12 studies that meet our inclusion criteria. To evaluate the overall prevalence of TDR, we extracted raw data and analyzed prevalence estimates using Stata SE. Estimates of mixed-effects were calculated by random-effects meta-analysis, and the I²statistics were used to estimate the heterogeneity of all included studies.Results: The Data of this meta-analysis come from 12 observational studies, covering 3558 ART-naive individuals with PHI, AHI or EHI. The overall prevalence of HIV-TDR is 9.8% (95% confidence interval (CI): 7.2%–12.3%, p<0.001). Prevalence of resistance by drug class is highest for the NNRTIs at 5.9% (95% CI: 3.1%–8.6%, p<0.001), followed by NRTIs 3.4% (95% CI: 1.8%–5.0%, p<0.001) and PIs 3.4% (95% CI: 2.7%–4.0%, p<0.001). The prevalence of TDR to INIs is 0.3% (95% CI: -0.1%-0.7%, p<0.001), which is the lowest among all antiretroviral drugs.Conclusion: The overall prevalence of TDR is high among AHI patients who have never received ART. This emphasizes the importance of baseline drug resistance testing for public health surveillance and guiding the choice of ART. In addition, the prevalence of TDR to NNRTIs is the highest, while the TDR to INIs is the lowest. This may guide the selection of clinical antiretroviral drugs.


Background
With the widespread application of antiretroviral therapy (ART) globally, HIV-related morbidity and mortality have fallen dramatically [1]. But as patients get more access of ART, the risk of transmitted drug resistance (TDR) has also increased [2][3][4]. Mistakes in HIV replication are the basis for virus to mutate and develop drug resistance [5], which can diminish the virological response to ART [6]. The relations between inadequate viral inhibition, poor treatment outcome and emergence of drug resistance are well understood [7]. TDR, also known as primary drug resistance, is de ned as the drug resistance caused by infection of drug-resistant mutant virus in newly diagnosed HIV patients with ART-naive. Pretreatment drug resistance (PDR) can be discovered in antiretroviral-naive people who begin receiving ART or people who have previous exposure to antiretroviral medicine and start or re-start ART now. In summary, PDR could have been transmitted during infection (e.g., TDR), or it may be acquired after being exposed to antiretroviral drugs (e.g., acquired drug resistance), or both.
TDR is a persistent public health problem that may affect ART at the population level [8]. People infected with HIV in the acute phase are highly contagious, and different kinds of drug resistance may exist before receiving ART. Mathematical models also have proposed that treatment failure in patients who have not undergone ART was most likely to be caused by the preexistence of resistant mutants [9]. It reminds us that patients with acute HIV infection (AHI) offer an occasion for real-time monitoring of TDR [2]. Previous studies have shown that the prevalence of TDR during AHI is higher than that in patients with chronic HIV infections [10]. Testing for TDR during AHI can improve the sensitivity of drug-resistant strains detection before the overgrowth of drug-sensitive viral quasi-species [10][11][12]. Baseline testing for TDR is routinely performed in the United States and Europe [13,14], but it is not the criterion in most resource-limited countries. Drug resistance mutations not only severely limit the options of treatment for new patients, but also accelerate the failure of treatment, and thus result in a waste of medical resources. The widespread of TDR could also undermine the planning effectiveness of national treatment efforts [6]. Therefore, routine detection of TDR is necessary to evaluate the emergence and spread of drug resistance mutations [15,16].
Our study mainly aims at the prevalence of TDR in AHI/ early HIV infection (EHI)/ primary HIV infection (PHI) patients who have not been given any antiretroviral drugs, to provide information on HIV therapy guidelines and governmental decisions on universal TDR testing, especially in some resource-limited places.

Search strategy and selection criteria
We conducted this meta-analysis of the studies published during January 1, 2008 to December 30, 2019, about the prevalence of TDR in ART-naive patients with AHI/ EHI/PHI, using the same search terms in PubMed, Embase, Web of science and the Cochrane Library (Table S1, Supplemental Digital Content). This research selects only original studies, excluding review articles and non-English studies. Notably, study selection criteria. The inclusion criteria only considered initial studies whose experiments must be done after January 1, 2000. We excluded studies of patients who had been exposed to antiretroviral drugs or those who were not in the acute phase of HIV infection. We excluded those studies that do not describe detailed information on drug resistance and patients' treatment involving ART, as well as ones that focus on patients with mother-to-child transmission and co-infection. We also excluded studies that were mixed with recent or chronic infections or unable to provide objective data.

Study selection and extraction
Two investigators (Chunxiang Guo and Yaxin WU) independently screened the titles and summaries, then evaluated the full text of the record and extracted the relevant data. Divergences were settled through consensus or a third reviewer (Yang Zhang)'s arbitration.
We extracted the following data from each study: nationality; sex; age; risk groups; study type; year of sample collection; number of participants; the number of patients with more than one drug resistance mutations (DRMs), with one or more nucleoside reverse transcriptase inhibitors (NRTIs) mutation, with one or more non-nucleoside reverse transcriptase inhibitors (NNRTIs) mutation, with one or more protease inhibitors (PIs) mutation, with one or more integrase inhibitors (INIs ) mutation, and with one or more Fusion enzyme inhibitors (FIs), and their respective mutation sites.

Risk of bias assessment
We used the Newcastle-Ottawa Quality Assessment Scale(NOS) [17] for the non-randomized trial and the observational study to evaluate the bias quality and risk of the included studies, and evaluated the methodological quality according to the participants' selection and hybrid control. A funnel plot and Egger's test were applied to detect any publication bias. A p-value of < 0.05 was considered statistically signi cant.

Data analysis
The main purpose of this meta-analysis is to describe the prevalence of TDR to NRTIs, NNRTIs, PIs and INIs in ART-naive patients with AHI, EHI, or PHI.
As for HIV genotype sequences, we focused on the mutations sites in Stanford University Drug Resistance Database [18] and the World Health Organization (WHO) surveillance drug resistance mutation list [19].
Statistical analysis was done in Stata SE. Heterogeneity was assessed using the I² statistic, when I²<25%, 25%-50%, and > 50%, the heterogeneity was low, medium and high, respectively. If I² is lower than 50%, we apply the xed model, otherwise, the random model. We set the con dence limits (CI) as 95% for each estimate. We calculated the proportion of speci c mutants after roughly pooling the numbers of individuals with any mutation and the number with particular mutations.

Study included and characteristics
We initially identi ed 1192 potential records from 4 electronic databases according to our search strategy. After deleting studies that do not meet our inclusion criteria, twelve full-length papers [1,2,10,[20][21][22][23][24][25][26][27][28] were analyzed, which represented 3558 participants with data on transmitted drug resistance-associated mutations (TDRAMs) (Fig. 1). Most of the study data came from voluntary counseling and testing institutions, blood donation centers, hospitals and pathology laboratories. Four of the seven countries involved in the study are developed countries. Characteristics of concerning studies were listed in Table 1.
Rare TDR mutations about PI-related minor genotypic mutations are L76V, N83D and L24I mutations, which are found in one person respectively (Fig. 5(C)). The funnel plot (Fig. S3 (D)) and Egger's test (P = 0.228) shows no signi cant publication bias.

Quality Assessment
The NOS [17] is used to evaluate the literature quality and bias risk of non-randomized controlled trials, including the study population, the comparability between groups and the measurement of results. The more the number of stars ( ), the better the quality is, and studies that rate 5 stars or more can be included in the analysis. The quality evaluation of the 12 included studies are all above 5 stars (Table S2). Participation bias may result from particular participant eligibility criteria in most studies. Participant retention was high overall. Publication bias, as shown earlier, generally speaking, shows no signi cant publication bias.

Discussion
We observed an overall TDR prevalence of 9.8% (95% CI: 7.2-12.3%) in this meta-analysis. Western developed countries have relatively high TDR prevalence due to the earlier popularization of ART. Among different categories of HIV drugs, patients show the highest resistance to NNRTIs, while shows lowest to the TDRAMs in INIs. This will help inform donors and policymakers on the urgent need to address drug resistance in AHI/ EHI/ PHI people without ART in an environment where high epidemic resistance has emerged.
At present, the antiretroviral regimen for treatment-naive patients generally consists of two NRTIs plus a third active agent selected from INIs, NNRTIs or boosted PIs with pharmacokinetic enhancers [13,14,29]. NRTIs plays a central role in ART and was widely used in the treatment of HIV infection. In the era of pre-combined ART, TDR was mainly targeted at NRTIs [23]. In our analysis, the TDR prevalence to NRTIs was second to NNRTIs. NRTI-related mutations, especially those that reduce the tness of the virus, can gradually disappear (such as M184V) [27]. Since NNRTIs were widely used, the prevalence of TDR to NNRTIs has increased over time. Our analysis indicates that the TDR of NNRTIs was the highest among the four types of inhibitors, and the prevalence is as high as 10% in developing countries. In our study, the most common NNRTIs-associated mutations are K103N/S, which can largely reduce the susceptibility or virological response, and develop resistance to NVP or EFV [18]. A single amino acid mutation can produce resistance to NVP or EFV, and the DRMs in NNRTIs are particularly important in predicting the effectiveness of rst-line regimens [20,30]. Besides, a few variants may persist and have a negative impact on the response to treatments [12,31]. But in some developing countries, NNRTs-based (e.g. EFV-, or rilpivirine-based) programs, are still the choice for some AIDS patients to begin their treatments [32]. Although the TDR prevalence to PIs was second only to NNRTIs, the clinical consequences caused by the increase of single PIs resistance mutations are very limited because of the high genetic barrier [27,33,34]. Nowadays, the two-drug regimens containing dolutegravir (DTG) + lamivudine (3TC) have become a relatively new scheme recommended as rst-line ART [14]. The existing data have shown that the TDR of INIs was the lowest in all kinds of inhibitors. Since 2014, INIs was listed as the preferred option for ART-naive patients due to their excellent e cacy and safety [35][36][37][38]. It is worth mentioning that although cases of drug resistance transmission of INIs have been reported, major INIs mutations are rarely found [21,22]. By this token, a PIbased and/or an INI-based combination over NNRTI-based regimens should reduce the risk of early failure [1]. At the same time, as more and more patients choose INIs-based regimens, we should also be prepared for a possible increase in the spread of INIs mutations [21].
The emergence of HIV drug resistance may impair the effectiveness of antiretroviral drugs [39] and further impact global HIV response and ART promotion [22]. According to current guidelines, drug resistance tests were recommended in all cases right after the diagnosis is made [13,14]. Detection of drug resistance mutations in AHI patients may be an important means to predict future drug resistance patterns [10] and help provide information on both global guidelines for the treatment and management of HIV.
As far as we know, this is the rst meta-analysis of TDR and mutations in ART-naive person with PHI, AHI, or EHI. We applied a systematic, comprehensive and quantitative method in the analysis and collected relatively comprehensive data and results of DRMs in acute phase of HIV infection.
There are also limitations in our study. Firstly, similar to most meta-analyses, the analyzed data were extracted from the literature, rather than the original data, which may lead to a small bias in data selection. Secondly, because there were relatively few tests for DRMs in ART-naive patients with PHI, AHI or EHI, the analysis only includes a small number of studies.
Thirdly, the data used in the study were mostly from urban and peri-urban areas, so the analysis mainly re ects the situation of DRMs in these areas. Despite these limitations, our meta-analysis can still provide references for ART-naive patients with AHI to choose antiretroviral regimens before starting treatment, especially in countries with NNRTI-based ones.

Conclusion
The moderate TDR prevalence detected among ART-naive patients with AHI/ EHI/ PHI in this systematic review and metaanalysis, highlights the signi cance of routine drug resistance detection before starting ART, especially in developing countries. The nding that a few NNRTI-associated (K103N/S and Y181C/I/V) DRMs were responsible for most cases of TDR suggests that the current available rst-line ART regimens containing EFV or NVP should be urgently amended. On the basis of these ndings, the direction of our follow-up works lies in expanding a variety of viral load measurements and further optimizing ART regimens.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests.