Survey of pretreatment HIV drug resistance and the genetic transmission networks among HIV-positive individuals in southwestern China, 2014–2020

Background Pretreatment drug 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-positive 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 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. Results In total, 3236 eligible HIV-positive individuals were included. The overall prevalence of PDR was 6.0% (194/3236). The PDR frequency to NNRTI (3.3%) was much higher than that of NRTI (1.7%, p < 0.001) and PI (1.2%, p < 0.001). A multivariate logistic regression analysis revealed that PDR was significantly higher among individuals aged 18–29 (adjusted odds ratio (aOR): 1.79, 95% CI 1.28–2.50) or 30–49 (aOR: 2.82, 95% CI 1.73–4.82), and harboring CRF08_BC (aOR: 3.23, 95% CI 1.58–6.59). A total of 1429 (43.8%) sequences were linked forming transmission clusters ranging in size from 2 to 119 individuals. Twenty-two individuals in 10 clusters had the same drug resistant mutations (DRMs), mostly to NNRTIs (50%, 5/10). Conclusions The overall prevalence of PDR was medium, numerous cases of the same DRMs among genetically linked individuals in networks further illustrated the importance of surveillance studies for mitigating PDR. Supplementary Information The online version contains supplementary material available at 10.1186/s12879-021-06847-5.


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 Open Access *Correspondence: ruanyuhua92@163.com 1 National Center for AIDS/STD Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing 102206, China Full list of author information is available at the end of the article has been identified 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 first-line therapy [4]. Therefore, timely monitoring of DRMs and the transmission of drug-resistant strains before initiation of ART can provide a scientific 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][7][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 first-line treatment, and if the percent of HIV-positive individuals harboring NNRTI resistance reaches above 10%, first-line treatments would need to be modified in these settings [13].
Guangxi in southwestern China, is one of the regions with the most severe HIV epidemic in China [14][15][16][17]. 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-positive 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-positive 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 individuals 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 persons 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, condomless sexual behavior in the past 3 months), and baseline CD4+ cell count before ART were collected from the National HIV/AIDS Comprehensive Response Information Management System [18].

Laboratory testing
Whole blood was sampled before the initiation of ART. For each person, CD4 cells were determined by the Alere PimaTM Analyzer (Abbott Laboratories, Germany). Frozen plasma (obtained by centrifuging whole blood) was isolated and sent on dry ice to the laboratory at National Center for AIDS/STD Control and Prevention (NCAIDS), Chinese Center for Disease Control and Prevention (CDC). RNA was extracted from 200 μl of plasma, amplified, and used to sequence the HIV pol region using an in-house Sanger sequencing protocol [8,19]. The threshold for mixture base calling was 20%. DRMs screening was identified and interpreted according to the Stanford University Genotypic Resistance Interpretation (https:// hivdb. stanf ord. edu/). Drug resistance to nucleoside reverse transcriptase inhibitors (NRTIs), non-nucleoside reverse transcriptase inhibitors (NNRTIs), and protease inhibitors (PIs) was defined as the detection of at least one antiretroviral virus drug in any drug class according to the WHO surveillance drug resistance guideline list [3]. Drug susceptibility was classified into four categories depending on mutation scores: susceptible-(< 15), low- (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29), intermediate- , and high-level (≥ 60) resistance. Any strain with a mutation score ≥ 15 was considered to be PDR.

Genetic network inference
The HIV genetic transmission network was inferred with HIV-TRACE [20], establishing putative transmission links between all sequences. We aligned HIV pol sequences to an HXB2 reference sequence (coordinates: 2253-3869) and calculated the pairwise genetic distances under the Tamura-Nei 93 (TN93) model [21]. Genetic networks with different genetic distances (range 0.1-1.5%) were established in order to find the most suitable genetic distance threshold that could identify the maximum number of clusters and links in the genetic network. However, Different genetic distance thresholds entertain different purposes. For example, a genetic distance of 0.5% would indicate that a strain would need a maximum time of 2-3 years in order to allow five different nucleotides to evolve, hence this threshold would be more aligned with rapid and more recent transmission events [22,23]. The aim of the study was to analyze transmission relationships associated with recent and rapid transmission, so 0.5% was preferentially selected as the optimal genetic distance threshold. To avoid potential deviation caused by convergent evolution, the genetic transmission network was reestablished after codons associated with drug resistance mutations in PR and RT were removed based on the main drug resistance mutation of HIV-1 recently updated [20,24], but the resulting network was unchanged. Molecular network was processed using the Cytoscape 3.5.2 software. Chord Diagram was created using the HIV-Trace that visualized trajectories of links between sequences in the molecular network.

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 χ 2 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. Specifically, all the variables that were statistically significant in the univariate analysis were included in the multivariate logistic regression model, then select the variables independently associated with PDR step by step based on goodness-of-fit tests (e.g., AIC). p < 0.05 was considered statistically significant. The 95% CI of the overall PDR prevalence is estimated using the distribution of the sample rate (p) approximating a normal distribution. The calculation formula is:  (Fig. 1). Within NRTI agents, PDR to 3TC and FTC were low at 0.1% (4/3236), yet these strains carried high-level resistance. Apart from that, low-level resistant strains were the main cases of ABC, AZT, TDF, D4T, and DDI, which were 0.4% (14/3236), 0.6% (18/3236), 0.3% (10/3236), 0.8% (26/3236), and 0.6% (18/3236), respectively. Within PI agents, the highest proportion of PDR was to NFV, with 15 (15/3236), 10 (10/3236), and 1(1/3236) in the low, middle and highlevel resistance, respectively. There was no HIV-1 strain resistant to DRV/r found in our study (Fig. 1). PDR did not significantly increase from 2014 to 2020 (p = 0.317, Mantel-Haenszel χ 2 linear trend test) (Additional file 1).

Factors associated with HIV PDR
Factors associated with HIV PDR are listed in CRF08_BC was also an important factor with an aOR of 3.23 (95% CI 1.58-6.59).

Genetic transmission networks
Genetic networks with different genetic thresholds (range 0.1-1.5%) were established, it is found that the molecular network with 0.5% gene threshold had a high number of clusters to get a high-resolution molecular network (Additional file 2). With the most suitable threshold genetic distance of 0.5%, a total of 1429 (43.8%) sequences were linked forming molecular transmission clusters ranging in size from 2 to 119 individuals (257 dyads, 99 clusters with 3 or more individuals). Fifty-nine sequences with PDR were found in  29 clusters (10 dyads, 19 clusters of three or more individuals), mostly to NNRTIs (17 of 29 clusters). Of note, the percentage of individuals with PDR belonging to a cluster was lower than of those without PDR (30.4% versus 44.7%, p < 0.001) (Additional file 3); nevertheless, the links between individuals with PDR versus without were no different within clustering individuals (2.00 versus 2.00, p = 0.345) (Additional file 4). There was a total of 4304 links in the molecular network, of which, with 19 (0.4%) links were PDR connected to PDR, 277 (6.4%) were PDR linked to No-PDR, and 4008 (93.1%) were No-PDR connected to No-PDR, the latter of which accounted for the largest proportion (Fig. 2). However, same DRMs yielded 10 transmission networks, distributed among CRF01_AE (10.3%, 3/29), CRF07_BC (3.4%, 1/29), CRF08_BC (17.2%, 5/29), and CRF55_01B (3.4%, 1/29) (Fig. 3). In the CRF01_AE subtype cluster, 29 cases of PDR individuals were enrolled, distributed among 18 clusters, and 12, 9, and 7 were resistant to NRTIs, NNRTIs, and PIs, respectively. In addition, there was one case of NNRTIs and NRTIs resistance with mutation sites V106I, V179D, and V75M. For the CRF07_BC subtype cluster, four cases of PDR individuals were found in two clusters. They currently reside in the same county. Among them, three cases in the same cluster had a history of non-marital heterosexual contact, and all carried Q58E resistant mutation sites. In the CRF08_BC subtype cluster, 22 cases of drug resistance entered the networks, distributed in seven clusters, mainly NNRTIs resistance (68.2%, 15/22) with the E138A mutation site. One case harbored both NRTI and NNRTI mutations, and the mutations sites were K103N and M46I. Three PDR persons in the CRF55_01B subtype belonged to the same network, with E138G and V179E mutations (Fig. 3).

Discussion
This cross-sectional study examined the prevalence of DRMs among 3236 ART-naive HIV-positive individuals in Qinzhou and obtained an overall PDR prevalence of 6.0% (194/3236, 95% CI 5.1-6.8%). According to the WHO definition of the prevalence of HIVDR, it is classified as low (< 5%), moderate (5-15%) and high (> 15%) [13]. The survey showed moderate (5-15%) rates of PDR 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 [25]. 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%) [26] and Liangshan of Sichuan (4.1 to 12.2%) [27,28] 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 [28]. The most common mutation of NRTIs is the K70K mutation, which is the same as a previous study [29]. The recommended firstline regimen was AZT or D4T + 3TC + NVP in China [30]. Since 2010, D4T has been gradually replaced by AZT or TDF. Currently, the first-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 first-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 [31]. These findings suggest that the current available first-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 influenced by many complex factors. This exploratory study found that age is an influential factor for PDR. Compared with people aged 50 and  [32,33]. Adherence to ART is likely lower in younger individuals, which increases their risk of harboring PDR. If these individuals engage in sexual activity, they could further transmit PDR strains to their sexual partners. However, whether it is due to increase sexual activity within certain individuals remains speculative and would require further study. Additionally, the frequency of mutations in the subtype CRF08_BC was significantly higher than that of CRF01_AE and CRF07_BC, which was consistent with the findings in Guangxi [11] and in Yunnan [34]. 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 persons 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 [35]. In addition, men who have sex with men and injection drug use were not found to be significant predictors of PDR in this study. In real-world research, the sample size of men who had sex with men or persons who inject drugs was low in this study, which could have reduced statistical power to determine whether these