Treatment effects of the differential first-line antiretroviral regimens among HIV/HBV coinfected patients in southwest China: an observational study

HIV with HBV co-infection can result in greater HIV-related immunosuppression, morbidity and mortality. Currently, there are few studies to evaluate direct treatment effects on mortality and attrition rates between first-line antiretroviral therapy (ART) based-on tenofovir (TDF) and/or lamivudine (3TC) in a real-world setting. We used Cox proportional hazard models to evaluate direct treatment effects of the first-line ART containing stavudine (d4T), azidothymidine (AZT) and TDF on death and attrition among HIV patients with HBV coinfection. A total of 3912 patients met study eligibility criteria. The overall mortality rate and attrition rate was 2.85 (95% CI: 2.55–3.16) and 8.87 (95% CI: 8.32–9.41) per 100 person-years, respectively. The ART containing TDF had a significantly lower risk of death [adjusted hazard ratio (AHR) = 0.58, 95% CI: 0.44–0.77] when compared to the ART containing d4T, but the risk of death was not significantly different when compared to the ART containing AZT (AHR = 0.91, 95% CI: 0.69–1.20). Patients with HIV/HBV coinfection receiving the ART containing TDF had significantly lower risk rates of attrition compared to those receiving the ART containing d4T (AHR = 0.72, 95% CI: 0.60–0.86) or AZT (AHR = 0.67, 95% CI: 0.58–0.77). Compared with the ART containing d4T, the ART containing AZT was significant and not significant associated with a lower risk of death and attrition, respectively. The ART containing TDF had significant effects on both of death and attrition among HIV patients with HBV coinfection.

there are an estimated 21 million HIV patients receiving antiretroviral therapy worldwide, enabling them the opportunity to live full and productive lives 18 . HIV and hepatitis B virus (HBV) infection is estimated at 35 million and 400 million worldwide respectively 19,20 . Due to HIV and HBV sharing similar pathways of transmission, many people living with HIV are co-infected with HBV, many of whom reside in Asian countries such as China 21,22 . HIV and HBV co-infection can greatly increase HIV-related immunosuppression, morbidity and mortality relative to either infection alone. The nucleoside reverse transcriptase inhibitors such as 3TC and TDF are effective for treatment of both HIV and HBV 20 . The World Health Organization (WHO) recommends initiation of ART for HIV/HBV co-infected patients and the NFATP recommended to switch the first-line regimen from didanosine to 3TC in 2008 23,24 . After that, the first-line regimens were AZT/d4T + 3TC + NVP in China. Patients with HBV treated only with 3TC quickly develop resistance so TDF is typically given in combination 25,26 . In 2012, TDF/AZT + 3TC + EFV/NVP regimens were introduced as the first-line treatments of the NFATP 24 . Patients with HIV and HBV coinfection treated with antiretroviral therapy containing TDF/3TC demonstrated the safety and effectiveness of the regimen 6,[27][28][29] . However, there are currently few studies to investigate associations between ART based-on TDF and/ or 3TC with direct treatment effects. We used the NFATP database from China to evaluate treatment effects of the ART containing D4T (d4T + 3TC + NVP), AZT (AZT + 3TC + EFV/NVP) and TDF (TDF + 3TC + EFV/NVP) on death and attrition among patients co-infected with HIV/HBV in an observational cohort study.

Results
Baseline characteristics of study patients. A total of 41,071 patients with HIV initiated combination antiretroviral therapy between 2010 and 2014 in Guangxi, China. Many patients (37,159) were excluded due to no HBV infection (25,639), no HBV testing results (11,056), lack of follow-up data (5 patients had no follow up besides baseline visit), being less than 18 years old (82), and not initiating a standard first-line treatment regimen (377). A total of 3,912 patients met study eligibility criteria and entered this observational cohort study analysis among patients with HIV/HBV coinfection. Table 1 presents baseline characteristics of the study patients. The proportion of study patients whose baseline age ≥50 years was 24.6%. 73.3% were male and 67.2% of study patients were married. The majority (84.3%) of study patients were infected through heterosexual intercourse, followed by homosexual intercourse (1.8%), intravenous drug use (11.8%) and unknown/other (2.2%). The prevalence of HCV infection was 13.5%. The proportion of study patients with CD4 cell count before ART ≤ 350 cells/μL were 88.3% and 44.9% of study patients were with WHO clinic stage III/IV before ART. The initial regimens used by study patients were: the ART containing D4T (d4T + 3TC + NVP) (14.8%), the ART containing AZT (AZT + 3TC + EFV/NVP) (29.2%) and the ART containing TDF (TDF + 3TC + EFV/NVP) (56.0%). The proportion patients who were currently using a first-line ART regimen was 90.2%.
Mortality rates and effects of initial ART regimen on death in study patients. Among Table 2). Table 3 presents adjusted effects of initial ART regimen on death in study patients. Comparing the ART containing D4T in the adjusted models, the ART containing AZT and TDF was significantly (AHR = 0.65, 95% CI: 0.48-0.89 and AHR = 0.58, 95% CI: 0.44-0.77) associated with a lower risk of death, respectively. In the adjusted model for sub-group analysis of effects of initial ART regimen on death, there was no significant differential between the ART containing AZT and the ART containing TDF.
Attrition rates and effects of initial ART regimen on attrition in study patients. Among all study patients, 979 attritions occurred, including 253 patients stopped ART and 726 patients were lost to follow-up. The overall attrition rate was 8.87 (95% CI: 8.32-9.41) per 100 person-years among all study patients. The attrition rate among all study patients initiating the ART containing D4T, AZT, and TDF was 9.31(95% CI: 7.98-10.64), 10.14(95% CI: 9.14-11.14) and 7.84 (95% CI: 7.12-8.57) per 100 person-years respectively (Table 4). Table 5 presents adjusted effects of initial ART regimen on attrition in study patients. Compared to the ART containing D4T in the adjusted models, the ART containing AZT and TDF was not significantly (AHR = 1.10, 95% CI: 0.92-1.31) and significantly (AHR = 0.72, 95% CI: 0.60-0.86) associated with higher and lower risk of attrition, respectively. Compared with the ART containing AZT in the adjusted model for sub-group analysis, the ART containing TDF (AHR = 0.67, 95% CI: 0.58-0.77) was significantly associated with lower attrition.

Discussion
In this observational cohort study, we found the overall mortality rate was 2.85 per 100 person-years among patients with HIV/HBV coinfection starting ART between 2010 and 2014 in Guangxi, China. The low mortality rate in our study was comparable to treatment effects seen in other developing and developed countries 7,21,[26][27][28][29][30] . This demonstrated a rapid but successful scale up of China's NFATP. Similarly, low mortality rates were also observed among HIV patients with HBV coinfection starting the ART containing D4T (d4T + 3TC + NVP) and AZT (AZT + 3TC + EFV/NVP) in our study. In China, initial ART regimens include 3TC as it is effective against both HIV and HBV. Our study findings also support reports that suggest the use of 3TC as a backbone of cART (combination ART) when treating patients with HIV who also test positive for HBsAg 31 .
Our large observational cohort study found that the ART containing TDF (TDF + 3TC + EFV/NVP) was significant associated with lower mortality risk among patients with HIV/HBV coinfection. Currently, the WHO recommends HBV/HIV co-infected patients start an ART regimen containing TDF and China's NFATP uses TDF + 3TC + EFV/NVP as the current initial first-line regimens 23,24 . Compared with the ART containing D4T in our study, the ART containing AZT was significant associated with lower death. This study results support China's NFATP to remove D4T due to side effects and efficacious 24 . Prior to this study, there was no observational data to support these recommendations in a real-world cohort.
Our study showed the overall attrition rate was 8.87 per 100 person-years among all study patients, which was higher than that in high-income settings, but lower than that in other low-income settings 34,35 . Compared with the ART containing D4T (d4T + 3TC + NVP) and AZT (AZT + 3TC + EFV/NVP), the ART containing TDF (TDF + 3TC + EFV/NVP) was associated with lower attrition risk among patients with HIV/HBV coinfection in this study. TDF has replaced both D4T and AZT in China, due to its more favorable toxicity profile, dosing, and cost-effectiveness 24 . The D4T regiment has salient side effects which can affect treatment outcomes. These findings also provide additional evidence to strongly support China's NFATP initiating TDF + 3TC + EFV/NVP among all HBV/HIV co-infected individuals since 2012, as advised by WHO guidelines 23,24 . Considering moderate attrition, it is essential that adherence is promoted amongst patients with HIV/HBV and physicians providing HIV care and ART receive comprehensive education and training to further improve clinical outcomes.
Our results should be interpreted within the limitations of this observational cohort study. First, selection bias of study participants in this cohort study may have been an issue, although we used the adjusted models to control for potential confounders. Due to our study did not include HIV diagnosed individuals but did not initiate ART, it is possible that individuals with poorer health-seeking behaviors were underrepresented in the study. Second, HBV infection was tested using an HBsAg ELISA which may not represent active HBV status. This definition could lead to overestimation of HBV infection, which may reduce statistical significance. Third, patients who were lost to follow-up might have different treatment outcomes compared to patients who remained in the study, such as high mortality and viral load rebound due to stop ART treatment. Fourth, due to this cohort study being conducted in Guangxi, study results might not be fully representative of China or other countries providing ART.
In conclusion, this is the first large-scale observational cohort study in a low-or middle-income countries to assess the effects of initial ART regimen on death and attrition in HIV patients with HBV coinfection. China aims to achieve the UNAIDS "90-90-90" target for significant reductions in HIV-related mortality and HIV new infection 36     the ART containing D4T, AZT and TDF (TDF + 3TC + EFV/NVP). We believe these findings demonstrating the utility of ART regimens containing TDF in a prospective observational cohort in China and will help optimize first-line antiretroviral therapy among patients with HIV/HBV coinfection in China and other countries.

Study design and study participants. This observational cohort study was conducted in the Guangxi
Zhuang Autonomous Region of southwest China. In 2016, Guangxi represented 10% of the total number of nationally reported HIV cases and this region has accumulated the third highest number of HIV cases reported in China. Heterosexual transmission was the main mode of transmission accounting for 93% of reported HIV cases in Guangxi 37 . Patients with HIV/HBV co-infection from the NFATP database were collected from 2010 to 2016, the characteristics of this database have been previously described 6,7,10,24 . Physicians administering the ART at the local hospitals manage case report forms at the time of initiating ART and follow-up at 0.5, 1, 2, 3 months, and every 3 months thereafter. The case report forms are uploaded into a web-based database hosted by China Centre for Disease Control (China CDC). Eligibility criteria for this study included: (1) HIV patients that were over 18 years old; (2) had positive HBV surface antigen results at the time of ART initiation; (3) initiated free first-line ART between 2010 and 2014; and (4) provided informed consent. The NFATP treatment criterion since 2010 requires at least one of the following: (1) WHO disease stage 3 or 4 (CD4 count of ≤350 cells/μL); (2) willingness to receive ART irrespective of other criterion, such as high CD4 count when initiating ART 24 . In China, the first-line regimens of AZT/d4T + 3TC + NVP were changed to TDF/AZT + 3TC + EFV/NVP in 2012, as recommended by WHO 24 . This study was approved by the institutional review board of the Guangxi Center for Disease Control and Prevention. All research methods in this study were carried out in accordance with the approved guidelines.
Data collection. The NFATP observational cohort study included baseline and follow-up characteristics.
Baseline characteristics included demographic variables (age, sex, marital status) and clinic variables (route of HIV transmission, HBV infection before ART, HCV infection before ART, CD4 cell count (cells/mm 3 ) before ART, WHO clinic stage before ART, initial ART regimen, current ART regimen, year of ART initiation). Follow-up characteristics were evaluated at every 3-months and included death or attrition. Variables collected at each follow-up included transferals to another clinic, cessation of ART, loss to follow-up, duration of ART, and survival status, which were obtained from hospital records or having clinic doctors' call family members to inquire about death. HBV infection was defined as positive serum samples for Hepatitis B surface antigen (HBsAg) which was tested using an HBV ELISA. Samples were tested for antibodies to HCV by ELISA 6 . Statistical analysis. We conducted a prospective follow-up study analysis. Time zero was defined as the date of ART initiation, and data was censored on June 30 th , 2016. Patients were censored if they were still alive or transferred to another clinic for care. Study endpoints included death and attrition. Attrition included lost to follow-up or withdrawal of ART as recorded in the HIV treatment database. Lost to follow-up was defined as missing more than 90 days after the last date seen in clinic, which was also defined as the date of withdrawal, as previous published studies 9,10 . Mortality and attrition rates were calculated based on Poisson distributions, and reported as deaths and attritions per 100 person-years of follow-up study, respectively.
In order to assess treatment effects of the differential first-line antiretroviral therapy, we classified the initial ART regimens as three groups: the first-line ART containing D4T (d4T + 3TC + NVP), the first-line ART containing AZT (AZT + 3TC + EFV/NVP) and the first-line ART containing TDF (TDF + 3TC + EFV/NVP). We used Cox proportional hazard models to evaluate treatment effects of initial ART regimen on death and attrition (lost to follow-up or withdrawal of ART), respectively. Competing risks for cause-specific hazard models were censored accordingly [9.10]. The adjusted models were used to control for potential bias. The following covariates of the adjusted model were included demographic variables (age, sex, marital status) and clinic variables (route of HIV infection, CD4 count before ART, WHO clinic stage before ART, year-initiated ART). And then, the adjusted model for sub-group analysis of death and attrition was used to compare the ART containing AZT with the ART containing TDF, respectively. A two-sided p-value less than 0.05 was regarded as statistically significant for all analyses. We did the analyses with Statistical Analysis System (SAS 9.1 for Windows; SAS Institute Inc., NC, USA).