Transmitted HIV Drug Resistance Among Individuals With Newly Diagnosed HIV Infection

A Multicenter Observational Study

Jingrong Ye; Mingqiang Hao; Hui Xing; Fujie Zhang; Hao Wu; Wei Lv; Tianjun Jiang; Yuncong Wang; Juan Wang; Yi Feng; Ruolei Xin; Ji Zeng; Shuai Zhao; Yinxiao Hao; Jing Chen; Yuhua Ruan; Xue Li; Yiming Shao; Hongyan Lu


AIDS. 2020;34(4):609-619. 

In This Article


Sampling Strategy

The Beijing HIV laboratory network (BHLN), established in 1986, is a collaborative network of laboratories involved in HIV diagnosis. It was authorized by the Beijing Municipal Commission of Health, and includes a central confirmatory laboratory in the Beijing Center for Disease Prevention and Control (CDC), four additional HIV confirmatory laboratories (DiTan, YouAn, Peking Union Medical College, and PLA General Hospital), and about 280 HIV screening laboratories. The collaboration maintains a biobank, which includes more than 50 000 samples from 25 648 individuals with HIV diagnosed in Beijing since 1986. BHLN takes part in maintaining the national HIV epidemiology database, which tracks everyone who receives a diagnosis of HIV infection in China and records the baseline CD4+ cell count of all individuals with newly diagnosed HIV infection. TDR has been monitored in China since 2006.[11] To ensure representativeness and feasibility, we designed a simple sampling strategy. In detail, however, we randomly selected 40% of the samples from all individuals with newly diagnosed HIV infection. In addition, we retrospectively included about 40% of samples that were stored in the biobank before routine genotyping was introduced in China.

Patient Inclusion and Data Collection

We included samples of individuals aged 18 years or older, with newly diagnosed HIV infection, and who were not pregnant. We excluded individuals who reported previous use of antiretroviral drugs for treatment or prophylaxis. We extracted baseline data on these individuals from the national HIV epidemiology database, including demographic characteristics, transmission risk group, and CD4+ cell count. ART coverage (which we defined as the proportion of all HIV-infected individuals receiving ART, with the total number of patients living with HIV as the denominator) was obtained from the National Annual Meeting on the Prevention and Treatment of AIDS and Hepatitis C of China (unpublished data, available upon request). For geographic location, we grouped individuals into 32 provinces according to the Hukou (see supplementary methods, We further categorized regions within each Hukou into five groups: first-tier cities, provincial capital, general city, county or town, and countryside or village, with the first four groups were regarded as urban, and the last group was regarded as rural. We established four sampling phases, according to the development of guidelines for NFATP in China: initial and scale-up ART, 2001–2008; standardization and scale-up ART, 2009–2012; rapid scale-up ART, 2013–2015; and treat-all, 2016–2017.[4]

Genotypic Resistance Analysis

We did population-based sequencing of HIV protease and codons 1–300 of reverse transcriptase on all specimens, using in-house methods.[11] We estimated the prevalence of TDR with the Stanford Calibrated Population Resistance (CPR) method, based on the 2009 WHO list of Surveillance of TDR Mutations (STDRM).[19]

Phylogenetic Analysis

We inferred HIV subtype by automated subtyping in Context-Based Modeling for Expeditious Typing (COMET)-HIV,[20] followed by maximum likelihood phylogenetic analysis of the pol sequences. To explore the relationship of the sequences of our study within the national epidemic, we merged our sequences with a nationally representative reference dataset. We downloaded all the pol sequences sampled in China with known sampling province and sampling year available in Los Alamos HIV sequence database(data available as of 26 July 2018) and used it as a nationally representative reference dataset (see supplementary methods, We then reconstructed an maximum likelihood phylogenetic tree with the merged dataset using the GTR+CAT nucleotide substitution model in FastTree 2.1 to check the relationship.[21]

Ethical Issues

The research ethics committee at the Beijing CDC approved the study. By law, consent was not required as these data were collected and analyzed in the course of routine public health surveillance.

Statistical Analysis

We compared categorical data with the χ 2 test and continuous data with one-way ANOVA wherever appropriate. We calculated the prevalence of sequences containing at least one TDR mutation and further specified for nucleoside reverse transcriptase inhibitors (NRTIs), non-NRTIs (NNRTIs), and protease inhibitors. We analyzed the potential risk factors for acquiring any TDR mutations using logistic regression. We assessed biologically plausible interactions in the multivariable model. Variables considered were sex, age (18–24, 25–44, 45–64, and ≥65 years), ethnicity, province, region of residence, HIV subtype, CD4+ cell count (<200, 200–349, 350–499, and ≥500 cells/μL), transmission risk group, ART coverage, and sampling phase. In the model, we included a binary response, indicating detection of any TDR mutation from each patient as an outcome. We analyzed all variables separately and entered those associated (P < 0.1) with the outcomes into the multivariable model. We expressed our results as odds ratios (ORs) with 95% confidence intervals (CIs) and two-sided P values, with a P value of less than 0.05 considered statistically significant. We did all analyses using R (version 3.5.0; R Foundation, Vienna, Austria). We used listwise deletion to handle missing data.

Definition of Recent Infection

To account for potential reversion of TDR mutations in the absence of drug pressure,[22] we repeated our analysis in the subset of patients who had been recently infected with HIV (infected for ≤1 year). We identified individuals as being recently infected based on the ambiguity score[23] (see supplementary methods,