Survey of Pretreatment HIV Drug Resistance and The Genetic Transmission Networks Among HIV-Infected Individuals in Southwestern China, 2014-2020

Pretreatment resistance (PDR) can limit the effectiveness of HIV antiretroviral therapy (ART). The aim of this study was to assess the prevalence of PDR among HIV-infected individuals that initiated antiretroviral therapy in 2014–2020 in southwestern China. Methods Consecutive cross-sectional surveys were conducted in Qinzhou, Guangxi. We obtained blood samples from individuals who were newly diagnosed with HIV in 2014–2020. PDR and genetic networks analyses were performed by HIV-1 pol sequences by using the Stanford HIV-database algorithm and HIV-TRACE, respectively. Univariate and multivariate logistic regression models were used to explore the potential factors associated with PDR.


Background
Since the establishment of China's national free HIV antiviral therapy in 2003, the case fatality rate of HIV infection has been effectively reduced and the life span of people living with HIV/AIDS has been prolonged [1,2]. However, with the increasing use of antiretroviral drugs, drug resistance has become an urgent problem. Pretreatment drug resistance (PDR) means that drug resistance has been identi ed by testing the resistance prior to antiviral therapy, including transmitted drug resistance (TDR) or primary resistance, or having received prior treatment, and then restarting antiviral therapy [3]. Studies found that the risk of virological failure within 12 months of treatment in people who developed PDR is 2-3 times higher, which will reduce the long-term effectiveness of rst-line therapy [4]. Therefore, timely monitoring of DRMs and the transmission of drug-resistant strains before initiation of ART can provide a scienti c basis for large-scale prevention and control program in China.
To address the challenges posed by drug-resistant viruses, WHO formulated a global strategy for the prevention of HIV drug resistance in 2012 [5]. Subsequently, it formulated the monitoring of drug resistant people who initiated antiretroviral therapy [3], which was recommended to monitor HIV resistance levels and factors associated with HIV drug resistance. National surveillance data showed that the prevalence of PDR was still at a low level in China [6-8], but the latest studies showed that PDR has reached a medium level in many parts of the country, such as Shanghai (17.4%, 55/317) [9], Tianjin (11.5%, 35/305) [10], Guangxi (7.21%, 83/1151) [11], and Beijing (6.12%, 57/932) [12]. According to the consolidated guidelines on the use of antiretroviral drugs, the proportion of NNRTIs resistance is much higher than that of NRTIs and PIs in countries with NNRTIs-based rst-line treatment, and the rate of NNRTIs is > 10%, and thus it is necessary to change the rst-line treatment drugs [13].
Guangxi in southwestern China, is one of the regions with the most severe HIV epidemic in China. However, little comprehensive data are available on PDR. In this study, we conducted a large sample survey of HIV drug resistance to assess the level of PDR in recent years (2014-2020). HIV-infected individuals had genotyping and DRMs monitoring prior to initiation of antiretroviral therapy. A genetic transmission network was constructed to explore PDR related transmission.

Study design and study population
This was a cross-sectional study to estimate the prevalence of PDR in HIV-infected individuals that initiated ART in the Prefecture of Qinzhou, one of the regions with the most severe HIV epidemic in Guangxi. The study design was done according to the 2014 WHO protocol note for PDR [3]. Eligibility criteria were as follows: HIV-1 patients newly diagnosed between January 1, 2014 and June 30, 2020, aged ≥ 18 years, signed the informed consent for PDR testing, and blood samples were successfully collected. Excluded from this study were patients who may have acquired drug resistance from previous antiretroviral drug exposure. After providing written informed consent, participants donated a single blood sample for HIV sequencing, HIV genotyping, and CD4 + T cell count assessment.

Data collection
Basic sociodemographic data (age, sex, ethnic, education, marital status, and occupation), behavioral characteristics (route of HIV infection, unprotected sexual behavior in the past three months), and baseline CD4 + cell count before ART were collected from the National HIV/AIDS Comprehensive Response Information Management System [14].

Genetic network inference
The HIV genetic transmission network was inferred with HIV-TRACE [16], establishing putative transmission links between all sequences. We aligned HIV pol sequences to an HXB2 reference sequence and calculated the pairwise genetic distances under the Tamura-Nei 93 (TN93) model [17]. Genetic networks with different genetic distances (range 0.1%-1.5%) were established in order to nd the most suitable genetic distance threshold that could identify the maximum number of clusters and links in the genetic network [18]. The genetic transmission network was reconstructed after removing all of the major DRMs from the sequences so that they would not impact the genetic distance comparison [19], but the resulting network was unchanged. For visualization and analysis, the network data were processed using the Cytoscape 3.5.2 software.

Statistical analysis
Statistical analyses were done in SAS V9.4 (SAS Institute Inc., Cary, NC, USA). Categorical variables are presented as number of cases and percentages, while continuous variables are expressed as the mean ± standard deviation (SD). Categorical variables which are presented as the number of cases and percentages were compared using the Pearson's χ² test or Fisher's exact test, while continuous variables are expressed as the mean ± standard deviation (SD) and compared using the non-parametric Mann-Whitney U-test and Kruskal-Wallis test, where appropriate. Univariate and multivariate logistic regression models were used to explore associations between PDR with demographic and clinical variables applied to each subregion. P < 0.05 was considered statistically signi cant.

Factors associated with HIV PDR
Factors associated with HIV PDR are listed in  subtype were enrolled in the network, with E138G and V179E mutations (Fig. 3).

Discussion
This cross-sectional study examined the prevalence of DRMs among 3236 ART-naive patients in Qinzhou and obtained an overall PDR prevalence of 6.0% (194/3236, 95% CI: 5.1%-6.8%). According to the WHO de nition of low, medium, and high levels of HIV-1 drug resistance (< 5%, 5-15% and > 15%) [20], PDR prevalence was at a medium epidemic level in the Qinzhou region. It was higher than the prevalence in Qinzhou in 2012-2013 (2.6%, 1/38), as well as in other regions of Guangxi [21]. From 2014 to 2020, the overall prevalence of PDR had no downward trend in Qinzhou, Guangxi. Similarly, PDR rates are on the rise in the provinces with the most severe HIV epidemic such as Guangxi in China, similar to Dehong of Yunnan (3.48 to 9.48%) [22] and Liangshan of Sichuan (4.1 to 12.2%) [23,24] from 2009 to 2017. In addition, the most common mutations of NNRTIs were E138A, K103N, and V179D mutations in this study, which was consistent with the study in a nationwide pilot survey of people with HIV/AIDS not receiving ART [24]. The most common mutation of NRTIs is the K70K mutation, which is the same as a previous study. [25]. The recommended rst-line regimen was AZT or D4T + 3TC + NVP in China [26]. Since 2010, D4T has been gradually replaced by AZT or TDF. Currently, the rst-line therapy is TDF or AZT + 3TC + EFV or NVP, which has effectively reduced acquired drug resistance. However, with the exception of D4T, the prevalence of NNRTI PDR was 3.3% (95% CI: 2.6-3.9) in this study, which is below the 10% threshold for changing the recommended rst-line antiretroviral therapy [13]. Similarly, some studies in the past have shown that PDR is primarily driven by resistance to NNRTI, with higher resistance to EFV, NVP, and/or RPV in particular [27]. These ndings suggest that the current available rst-line ART regimens containing D4T, EFV, and/or NVP and/or RPV need to be revised. In addition, it is recommended that there be drug resistance testing and viral load measurements prior to ART initiation.
PDR may be in uenced by many complex factors. This exploratory study found that age is an in uential factor for PDR. Compared with people aged 50 and above, young people aged 18-29 are more likely to have pre-treatment drug resistance, in line with studies in Jiangsu Province, Shandong Province, Guangxi, and Vietnam [28,29]. The ndings suggest that younger people who have an active sexual life, advocate individuality, and poor adherence to medication are more likely to transmit resistant strains to a newly infected individual. Additionally, the frequency of mutations in the subtype CRF08_BC was signi cantly higher than that of CRF01_AE and CRF07_BC, which was consistent with the ndings in Guangxi [11] and in Yunnan [30]. It has been suggested that the subtype CRF08_BC strain is more prone to base mismatches at certain sites during replication, leading to higher mutation rates, and patients carrying the CRF08_BC virus with mutations E138G, M184I, Y181C, Y188C, L100I, and may be highly resistant to antiretroviral drugs including NVP, EFV, or 3TC [31].
In this study, PDR did not lead to an increase in the clustering rate, as did the ndings in 13 provinces or cities in China (include high and moderate prevalence regions) [24], Liangshan Prefecture in Sichuan [23] and Shijiazhuang in Hebei [32]. Studies have suggested that the transmission capacity of resistant strains is lower than that of non-resistant strains [33,34]. On the other hand, some of the newly diagnosed HIV-infected patients included in this study were not recently infected. They had not been treated with drugs after infection, so that the non-resistant strains became the dominant strains in the patients and drug-resistant mutations may not have been detected [35,36]. However, there were 10 clusters containing DRMs in the genetic network of this study, and the size of clusters tend to expand with the reporting time. This result suggested that PDR may be transmitted among high-risk groups in the future, and that interventions in these populations are necessary to prevent the spread of drug-resistant strains in the region.
Our study has limitations. The sample transmission categories were based on self-reported information, and we could not con rm the authenticity of the data. We plan to conduct a more detailed study design to obtain accurate epidemiological survey data. In addition, Sanger sequencing can only detect minority drug-resistant strains at a 15%-20% frequency of HIV viral populations of patients and PDR was underestimated in this study [37]. In the future, we hope that next-generation sequencing can be used to identify HIV drug resistant variants at frequencies as low as 0.4%.
In summary, large-scale drug resistance surveillance was carried out on HIV-infected individuals that initiated ART in Qinzhou, which has provided insights into the PDR prevalence, in uencing factors, and potential transmission relationships of drug-resistant strains in the region. These ndings indicate that there is an urgent need for surveillance programs of HIVDR and routine drug resistance testing in the clinical management of patients. For ART-naïve patients, the results of drug resistance monitoring could guide dosing regimens to improve the therapeutic effect. Also, behavioral intervention and traceability investigation can reduce or even block the continuous transmission of drug-resistant strains. For national institutions, large-scale studies can help develop national guidelines for ART.

Conclusions
In conclusion, large-scale drug resistance surveillance showed that the overall prevalence of PDR is moderate in Qinzhou. Pretreatment drug resistance did not lead to an increase in the clustering rate.