Drug Resistance and Virological Failure among HIV-Infected Patients after a Decade of Antiretroviral Treatment Expansion in Eight Provinces of China

Background China’s National Free Antiretroviral Treatment Program (NFATP) has substantially increased the survival rate since 2002. However, the emergence of HIV drug resistance (HIVDR) limits the durability and effectiveness of antiretroviral treatment (ART) in at risk patients. Method A cross-sectional survey was conducted among patients having received a median of 13.9 months of ART in eight provinces in China. Demographic and clinical information was collected, and venous blood was sampled for CD4 cell counts, measurement of the HIV viral load (VL), and HIV drug resistance (HIVDR) genotyping. Possible risk factors for HIVDR were analyzed by the logistic regression model. Results The study included 765 patients. Among them, 65 patients (8.5%) had virological failure (VLF) defined as ≥1,000 copies/ml. Among the individuals with VLF, 64 were successful genotyped, and of these, 33 had one or more HIVDR mutations. The prevalence of HIVDR mutations among patients receiving first-line ART was 4.3% (33/765). All of the patients with HIVDR mutations were resistant to non-nucleoside transcriptase inhibitors, 81.8% were resistant to nucleoside reverse transcriptase inhibitors, and only 3% had mutations that caused resistance to protease inhibitors. Having lower ratios of drug intake in the past month and dwelling in two southwestern provinces were factors independently associated with the emergence of HIVDR. Conclusion Most patients receiving first-line ART treatment achieved sound virological and immunological outcomes. However, poor adherence is still a key problem, which has led to the high rate of HIVDR. It was notable that the proportion of drug resistance widely varied among the provinces. More studies are needed to focus on adherence.


Introduction
Since its development in 1996, combination antiretroviral therapy (cART) has significantly improved the quality of life of HIV-infected persons and dramatically lowered their morbidity and mortality [1][2][3][4][5]. It has become widely available in most resource-limited or developing countries since the WHO launched the "3 by 5" initiative in 2003. It was reported at the world AIDS day 2015 that 16 million people were receiving antiretroviral treatment and 7.8 million HIV-related deaths had been averted between 2000 and 2015.
In China, the National Free Antiretroviral Treatment Program (NFATP) was begun in 2002, after a pilot study [6]. As a public health approached ART program, NFATP has proved to have efficiently reduced mortality among HIV-infected Chinese patients [7][8][9][10][11][12]. The development of NFATP was in three phases. The first phase was from 2002 to 2005, the second phase,also called as the first scale-up phase, was between 2005 and 2007, and the third phase which was the further scale-up and standardization phase was from 2008 onwards [13]. Significant policy changes in the third phase included scale-up HIV testing among key populations and immediate initiation of ART in China. By the end of 2014, more than 363,000 patients in China had received ART [14]. However, concerns for the emergence of drug resistance grew during the rapid ART expansion in China. In this study, we conducted a survey of acquired HIVDR based on the WHO HIVDR surveillance protocol in order to assess the level of virological suppression and drug resistance during these years in China. Our findings would provide valuable implications for good practice of planning treatments for all people living with HIV.

Study design and study population
We conducted a cross-sectional survey in eight provinces or cities of China: Beijing, Jilin, Hunan, Guangxi, Sichuan, Guizhou, Yunnan and Xinjiang Province. The survey protocol was taken from the WHO recommended cross-sectional survey on acquired HIVDR in adult patients receiving ART. Patients included were 18 years or older, had begun free ART treatment in 2013, and had received first-line ART for 9-18 months at enrollment. Eligible patients were enrolled at routine clinic visits in 2014. All participants provided written informed consent.

Ethics approval
Institutional review board approval was granted by National Center for AIDS/STD Control and Prevention (NCAIDS), Chinese Center for Disease Control and Prevention (China CDC).

Data collection
An interview-administered questionnaire (S1 File) was used for data collection. The questionnaire was administered face to face by trained local health staff in a private room. Data on demographic characteristics, ART treatment, and medicine adherence were collected during the interview.

Laboratory tests
Blood specimens were collected after the interview. CD4+ T cells were quantified using flow cytometry at local CDCs within 12 hours. Plasma was isolated and sent under cold chain to the laboratory at NCAIDS, China CDC where the HIV viral load was measured. Viral suppression was defined as an HIV RNA level <1000 copies/ml. In samples with a viral load !1000 copies/ ml, HIV drug resistance genotyping was performed at the NCAIDS laboratory by using an inhouse method [15,16]. A drug resistance mutation was identified and interpreted by using the algorithm of the Stanford HIV Drug Resistance Database (http://hivdb.stanford.edu/pages/ algs/sierra_sequence.html). HIV drug resistance mutations were defined as those conferring low-, intermediate, or high-level resistance [17,18].

Statistical analysis
All questionnaire data were double-entered using Epidata 3.1 (The Epidata Association Odense, Denmark). Statistical Analyses (S1 Table) were performed using SAS V9.4 (SAS Institute Inc, Cary, North Carolina, USA). Univariate logistic regression models were constructed to explore factors associated with drug resistance. A stepwise multivariate logistic regression model was constructed to select the variables that were independently associated with drug resistance. A P value <0.05 was considered statistically significant, and all tests were twosided.

Demographic characteristics
This survey included 765 patients having received ART for 9-18 months (

Immunological and virological outcomes
Among the patients, the proportion of a CD4 count of 0-199, 200-349, and !350 cells/ul before ART were 45.4%, 37.4%, and 17.2%, respectively. After 13.9 months of treatment, the proportion of a CD4 count of !350 cells/ul was increased to 37.4%. The median CD4 count before ART increased from 222 (IQR, 215) cells/ul to 303 (IQR, 258) cells/ul at the time of the survey. The great majority of patients (700/765, 91.5%) had a plasma HIV viral load <1000 copies/ml. Among the individuals with VLF, 64 were successful genotyped, and of these, 33 had one or more HIVDR mutations. However, the proportion of viral load failure among patients from Sichuan and Guizhou were 27% and 24%, respectively, which was higher than patients from other provinces.

HIV drug resistance and subtype
Among the 33 patients identified with HIVDR mutations, all patients harbored HIV-1 strains resistant to non-nucleoside reverse transcriptase inhibitors (NNRTIs) (

Patient characteristics associated with HIV drug resistance
The risk factors for HIVDR that were significant in the univariate logistic analysis were included in the multivariate logistic regression (Table 3). According to the univariate logistic regression model, five potential factors correlated with HIV drug resistance. In the multivariate model, the following two factors were independently correlated with HIVDR: the rate of HIVDR among patients with <90% of drug intake in the past month were 6.0 folds higher than in patients with !90% of drug intake (95%CI: 1.7-20.7; P = 0.005), and Sichuan and Guizhou Provinces were 7.3 times higher than the other provinces (95%CI: 3.6-15.2; P<0.0001).

Discussion
In this study, we analyzed HIVDR data in order to evaluate the prevalence and risk factors among 765 patients who were undergoing treatment. Our findings showed that 91.5% (700/ 765) achieved virological suppression (VL<1000 copies/ml) after 13.9 months of ART, which is better than the outcomes of previous surveillance studies in China [19][20][21]. 33 patients (4.3%) had verified HIVDR which is slightly lower than other countries where national free ART is available. The prevalence of HIVDR in Cameroon and Namibia were 5.3% and 5%   respectively [22,23] but significantly higher than Malawi [24] of 3.4%. Our results suggest that China's free ART program is providing high quality care to HIV/AIDS patients. There are several reasons to explain the low virological failure and drug resistance among HIV-infected patients receiving first-line ART. First, medical care is accessible at many levels of the health systems including provincial, prefecture, and county hospitals. Most care is provided at the community level and through outreach, with telephone calls or home visitation. Second, all doctors and health staff involved in providing ART and care management receive additional training [25].
Although our study showed that China has met the WHO target for 90% of patients having their viral load suppressed, as HIV/AIDS patients live longer and are on ART for life, the number of patients with drug resistance is likely to increase. A combination of strategies is required to combat drug resistance. New medicines that can more robustly cope with drug resistance mutants are needed, especially for those with common mutations such as K103N and M184I/ V [19,26,27]. M184I/V confers resistance to lamivudine, which is often the first mutation to develop in patients receiving partially suppressive triple combination therapy including lamivudine [28]. K103N is one of the most frequent mutations conferring resistance to most available NNRTIs [29].
Factors independently associated with the incidence of HIVDR were: the ratio of drug intake in the past month; and place of residence. The first factor reflected that adherence was a direct factor causing HIVDR, with 21(2.7%) patients having reported to have lower than 90% of drug intake in the past month. Similar findings had been reported in our previous studies in China [19,27,30], where poor adherence clearly leads to the occurrence of HIVDR. Good adherence can suppress plasma HIV RNA and utilize the optimum effectiveness of the ART therapy. Several studies have focused on strategies to improve adherence, including social support [31,32], behavioral interventions [33], contingency management strategies [34], directly administered antiretroviral therapy (DAART) [35], and technological interventions [36]. The Chinese strategy focuses on education and counseling to improve the adherence of patients; but a comprehensive strategy using some of the other interventions is needed.
The reasons why patients from Sichuan and Guizhou had worse outcomes compared to patients from other provinces was unclear. We found, however, that the composition of patients from this population differed compared to patients from other provinces. Their differential risk for drug resistance may have been mediated through factors influencing adherence such as education, economic level, healthcare providers support, and adherence to ART. 78.2% of the Sichuan and Guizhou patients had received middle school education or lower compared with 69.5% in other provinces, 32.7% of patients in Sichuan and Guizhou were poor versus 22.3% s in other provinces, 59.1% of patients in Sichuan and Guizhou get support from healthcare providers compared with 89.9% in other provinces, and 14.5% in Sichuan and Guizhou had a poor adherence to ART compared with 7.8% in other provinces. There may, however, have been unmeasured confounding variables that led to this observation. Observations about drug resistance found on the population level, may not apply to individual patients.  Table 3. Factors associated with HIV drug resistance among patients receiving ART in China. Future studies should explore health systems and individual level differences to better elucidate why resistance was greater in Sichuan and Guizhou provinces than other provinces. We also found that regimens with or without TDF showed the same results in causing HIVDR, which needs further investigation. TDF is preferred to its predecessors AZT and d4T in the ART program because of its better safety profile [37] which has been recommended by WHO for HIV first-line treatment. Researchers found that patients on TDF-based first-line This study has some limitations. First, this was a cross-sectional study and patients who terminated treatment (due to adverse reactions, loss to follow-up, or death) would not have been sampled, which may have led to overestimated treatment effectiveness. Second, adherence was assessed by self-reporting of having missed doses in the past month, which may not reflect the true adherence. Finally, we found the drug resistance difference between Guizhou, Sichuan, and other provinces. Explaining the differences focused on the socioeconomic status of the inhabitants, but in fact transmitted drug resistance and possible presence of transmission clusters with HIV drug resistant variants may differ in different provinces which we do not know.

Variable Number Drug resistance, N (%) Crude OR(95%CI) P-value Adjusted OR(95%CI) P-value
In conclusion, a representative national sample of HIVDR surveillance across China demonstrated excellent virological and immunological outcomes at 9-18 months among patients receiving first-line ART treatment. However, poor adherence to treatment is still a key problem regardless of the efforts on the regimens, which has led to the high rate of HIVDR. Drug resistance widely varies among provinces. More research needs to focus on the adherence of patients and long-term studies monitoring drug resistance should be completed in some select cases.
Supporting Information S1 File. Questionnaire of this research. (DOC) S1