Transmitted Drug Resistance in Antiretroviral Therapy-Naive Persons With Acute/Early/Primary HIV Infection: A Systematic Review and Meta-Analysis

Background: The widespread use of antiretroviral therapy (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 primary HIV infection (PHI)/AHI/early HIV infection (EHI) and to describe the critical drug-resistant mutations. Methods: We systematically searched the literature between January 1, 2008, and April 30, 2021, in PubMed, Web of Science, Embase, and the Cochrane Library. To evaluate the overall prevalence of TDR, we extracted raw data and analyzed prevalence estimates using Stata SE. Results: The data of this meta-analysis come from 12 observational studies, covering 3,558 ART-naive individuals with PHI, AHI, or EHI. The overall prevalence of HIV-TDR is 9.3% (95% CI: 6.8%–11.8%, I2 = 81.1%, in 11 studies). The prevalence of resistance by drug class is the highest for the nonnucleoside reverse transcriptase inhibitors (NNRTIs) at 5.7% (95% CI: 2.9%–8.5%, I2 = 96.6%, in 11 studies), followed by nucleoside reverse transcriptase inhibitors (NRTIs) at 3.4% (95% CI: 1.8%–5.0%, I2 = 86.3%, in 10 studies) and protease inhibitors (PIs) at 3.3% (95% CI: 2.7%–3.9%, I2 = 15.6%, in 10 studies). The prevalence of TDR to integrase inhibitors (INIs) is 0.3% (95% CI: 0.1%–0.7%, I2 = 95.9%, in three studies), which is the lowest among all antiretroviral drugs. Conclusion: The overall prevalence of TDR is at a moderate level 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.


INTRODUCTION
With the widespread application of antiretroviral therapy (ART) globally, HIV-related morbidity and mortality have fallen dramatically (Visseaux et al., 2019). According to current guidelines, an antiretroviral regimen for a treatment-naive patients generally consists of three active agents-two nucleoside reverse transcriptase inhibitors (NRTIs) (e.g., abacavir/lamivudine) and an antiretroviral drug selected from integrase inhibitors (INIs) (e.g., dolutegravir), non-NRTI (NNRTI) (e.g., efavirenz (EFV)), or boosted protease inhibitor (PI) (e.g., lopinavir) with pharmacokinetic enhancer (World Health Organization, 2016; Department of Health and Human Services, 2019; European AIDS Clinical Society (EACS), 2019). INIs are included as part of initial HIV therapy in most treatment guidelines (Department of Health and Human Services, 2019), and NNRTI-based regimens like EFV-or rilpivirine based regimens may also be options for some patients especially in developing countries to initiate therapy (AIDS and Hepatitis C Professional Group, 2018).
But as patients have more access of ART, the risk of transmitted drug resistance (TDR) has also increased (Hamers et al., 2011;Gupta et al., 2012;Colby et al., 2016). Mistakes in HIV replication are the basis for the virus to mutate and develop drug resistance (Abram et al., 2014), which can diminish the virological response to ART (Wittkop et al., 2011). The relations between inadequate viral inhibition, poor treatment outcome, and emergence of drug resistance are well understood (Bennett et al., 2008). TDR, also known as primary drug resistance, occurs when an uninfected and ARTnaive person is infected with a drug-resistant strain of HIV from someone with HIV drug-resistance mutations (DRMs). Pretreatment drug resistance (PDR) can be discovered in antiretroviral-naive people or people who start ART or had previous exposure (i.e., re-start) treatment. 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 (Gupta et al., 2018).
TDR is a persistent public health problem that may affect ART at the population level (Bansi et al., 2010). 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 (Ribeiro and Bonhoeffer, 2000). It reminds us that patients with acute HIV infection (AHI) offer an occasion for real-time monitoring of TDR (Colby et al., 2016). Previous studies have shown that the prevalence of TDR during AHI is higher than that in patients with chronic HIV infections (Yanik et al., 2012). Testing for TDR during AHI can improve the sensitivity of drugresistant strain detection (Jain et al., 2011;Yanik et al., 2012;Castro et al., 2013). Baseline testing for TDR is routinely performed in the United States and Europe (Department of Health and Human Services, 2019, European AIDS Clinical Society (EACS), 2019), but it is not the criterion in most resource-limited countries. DRMs 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 (Wittkop et al., 2011). Therefore, routine detection of TDR is necessary to evaluate the emergence and spread of DRMs (Hedt et al., 2011;Jordan et al., 2012).
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 April 30, 2021, 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 (Supplementary Table S1, Supplemental Digital Content). The selection criteria of the study are as follows: Firstly, we only include original research (e.g., cohort studies, prospective studies, case-control studies, and randomized controlled trials) on adults (aged >15 years), excluding review articles and non-English studies. Secondly, 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 the ones that focus on patients with mother-to-child transmission and coinfection. We also excluded studies that were mixed with recent or chronic infections or unable to provide objective data.

Study Selection and Extraction
Two investigators (CG and YW) 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's (YZ's) arbitration.
We extracted the following data from each study: country; sex; age; risk groups; study type; year of sample collection; number of participants; and the number of patients with more than one DRMs, with one or more NRTI mutation, with one or more NNRTI mutation, with one or more PI mutation, with one or more INI mutation, and with one or more Fusion enzyme inhibitors and their respective mutation sites.

Risk of Bias Assessment
We used the Newcastle-Ottawa Quality Assessment Scale (NOS) for the non-randomized trial and the observational study to evaluate the bias quality and risk of the included studies, including the study population, the comparability between groups, and the measurement of results (Wells et al., 2014). The more the number of stars (q), the better the quality is, and studies rated five stars or more can be included in the analysis. A funnel plot and Egger's test were applied to detect any publication bias. A p-value of <0.05 was considered statistically significant.

Data Analysis
The main purpose of this meta-analysis is to describe the prevalence of TDR to NRTIs, NNRTIs, PIs, and INIs in ARTnaive patients with AHI, EHI, or PHI.
As for HIV genotype sequences, we focused on the mutations sites in Stanford University Drug Resistance Database and the WHO surveillance drug resistance mutation list (Bennett et al., 2009;Stanford, 2020). The WHO classifies the level of TDR as low (<5%), moderate (5%-15%), and high (>15%) .
Statistical analysis was done in Stata SE. Heterogeneity was assessed using the I 2 statistic; when I 2 <25%, 25%-50%, and >50%, the heterogeneity was low, medium, and high, respectively. If I 2 is lower than 50%, we apply the fixed model; otherwise, the random model. Subgroup analysis was conducted according to study country to assess the heterogeneity between studies. We set the confidence interval as 95% for each estimate. We calculated the proportion of specific mutants after crudely pooling the numbers of individuals with any mutation and the number with particular mutations (Gupta et al., 2018).

Study Included and Characteristics
We initially identified 1,238 potential records from four electronic databases according to our search strategy. After deleting studies that do not meet our inclusion criteria, 12 fulllength papers were analyzed, which represented 3,558 participants with data on TDR-associated mutations (TDRAMs) ( Figure 1) (Kim et al., 2008;Castor et al., 2012;Yanik et al., 2012;Dai et al., 2014;Ambrosioni et al., 2015;Ananworanich et al., 2015;Stekler et al., 2015;Zhao et al., 2015;Colby et al., 2016;Panichsillapakit et al., 2016;Rutstein et al., 2019;Visseaux et al., 2019). One of the literatures included in this study was a cross-sectional survey, and 11 were cohort study. The sampling span was from 2000 to 2016, and there were eight studies with a sample size of more than 100. 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. The characteristics of concerning studies are listed in Table 1.

The Prevalence of Transmitted Drug
Resistance and Transmitted Drug Resistance-Associated Mutations

Quality Assessment
The NOS is used to evaluate the literature quality and bias risk of non-randomized controlled trials (Wells et al., 2014). The quality evaluation of the 12 included studies is all above five stars (Supplementary Table S2). Participation bias may result from particular participant eligibility criteria in most studies. Participant retention was high overall. As shown earlier, generally speaking, no significant publication bias was shown.

DISCUSSION
We observed an overall TDR prevalence of 9.3% (95% CI: 6.8%-11.8%) in this meta-analysis. Subgroup analysis shows that the prevalence of TDR in developed countries is higher than that in developing countries, and both at a moderate level . Among different categories of HIV drugs, patients show the highest resistance to NNRTIs, while they show the 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 (Department of Health and Human Services, 2019; European AIDS Clinical Society (EACS), 2019; World Health Organization, 2016). NRTIs play a central role in ART and are widely used in the treatment of HIV infection. In the era of pre-combined ART, TDR was mainly targeted at NRTIs (Panichsillapakit et al., 2016). In our analysis, the TDR prevalence to NRTIs was second to that to NNRTIs. NRTI-related mutations, especially those that reduce the fitness of the virus, can gradually disappear (such as M184V) (Ambrosioni et al., 2015). Since NNRTIs were widely used, the prevalence of TDR to NNRTIs has increased over time (Panichsillapakit et al., 2016). Our analysis indicates that the TDR of NNRTIs was the highest among the four types of inhibitors. And the most common NNRTI-associated mutations are K103N/S, which can largely reduce the susceptibility or virological response and develop resistance to NVP or EFV (Stanford, 2020). 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 first-line regimens (Zhao et al., 2015;Avila-Rios et al., 2016). Besides, a few variants may persist and have a negative impact on the response to treatments (Jain et al., 2011;Li et al., 2011). But in some developing countries, NNRTI-based (e.g., EFV-or rilpivirinebased) programs are still the choice for some persons infected with HIV to begin their treatments (AIDS and Hepatitis C Professional Group, 2018). The TDR prevalence to PIs was lower than that to NNRTIs and NRTIs, and the clinical consequences caused by the increase of single PI resistance mutations are very limited because of the high genetic barrier (Rhee et al., 2010;Wensing et al., 2010;Ambrosioni et al., 2015). Nowadays, the two-drug regimens containing dolutegravir + lamivudine have become a relatively new scheme recommended as first-line ART (Department of Health and Human Services, 2019). The existing data have shown that the TDR of INIs was the lowest in all kinds of inhibitors. Since 2014, INIs were listed as the preferred option for ART-naive patients due to their excellent efficacy and safety (Walmsley et al., 2013;Clotet et al., 2014;Lennox et al., 2014;Squires et al., 2016). It is worth mentioning that although cases of drug resistance transmission of INIs have been reported, major INI mutations are rarely found (Stekler et al., 2015;Rutstein et al., 2019). By this token, a PI-based and/or an INI-based combination over NNRTIbased regimens should reduce the risk of early failure (Visseaux et al., 2019). At the same time, as more and more patients choose INI-based regimens, we should also be prepared for a possible increase in the spread of INI mutations (Stekler et al., 2015). The emergence of HIV drug resistance may impair the effectiveness of antiretroviral drugs and further impact global HIV response and ART promotion (Zhao et al., 2015;Clutter et al., 2016). According to current guidelines, drug resistance tests were recommended in all cases right after the diagnosis is made (Department of Health and Human Services, 2019; European AIDS Clinical Society (EACS), 2019). Detection of DRMs in AHI patients may be an important means to predict future drug resistance patterns and help provide information on both global guidelines for the treatment and management of HIV (Yanik et al., 2012).
As far as we know, this is the first 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 the 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 reflects 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 meta-analysis highlights the significance of routine drug resistance detection before starting ART in both developed and developing countries. The finding that a few NNRTIassociated (K103N/S and Y181C/I/V) DRMs were responsible for most cases of TDR suggests that the current available first-line ART regimens containing EFV or NVP should be urgently amended. On the basis of these findings, the direction of our follow-up works lies in expanding a variety of viral load measurements and further optimizing ART regimens.

AUTHOR CONTRIBUTIONS
XH and GC contributed to the study conception as well as design. CG and YW collected the data, performed the statistical analysis, and drafted the manuscript. TZ, XL, AL, and MG conducted manuscript redaction. YZ, HW, XH, and GC made critical changes to the manuscript. All authors read and approved the final published manuscript.

FUNDING
This study was supported by the National Science and Technology Major Project of China during the 13th Five-year plan period (2017ZX10201101), the Beijing Excellent Talent Plan (2018000021223ZK04), the Beijing Talent Project in the New Millennium (2020A35) and the Beijing Hospitals Authority "Peak Climbing" Planning (DFL20191701).