The evaluation of antiretroviral treatment in HIV and hepatitis B virus (HBV) co-infected persons in Beijing, China, 2010-2018: a retrospective cohort study

There is limited long-term data on the effect of ART on clinical outcomes in HIV/HBV patients in China. The objective of this study was to understand the ART treatment effect and the factors associated with the loss to follow-up or death of HIV/HBV coinfected patients in the city of Beijing, China. Methods: This study examined clinical indicators for HIV mono-infected or HIV/HBV co-infected patients from Jan. 2010 to Sep. 2018. Included patients were followed for a mean duration of 34.5 months after ART. Covariance analysis for repeated measures was used to analyze the changes of clinical indicators; multivariate logistic regression and Cox model were used to analyze the influencing factors of the abnormal incidence of clinical indicators and the lost to follow-up or death in HIV patients. Results: A total of 841 HIV/HBV coinfected patients and 2000 HIV patients were analyzed. Adherence was estimated to be 93% in all patients. At baseline, ALT, AST, OIs and APRI≥0.5 in HIV/HBV patients were higher, while the CD4 and CD4/CD8 ratio were lower. After ART treatment, the rate of APRI≥0.5 (4.4%) were still higher in HIV/HBV patients and HBV co-infection affect the prevalence of APRI≥0.5 (OR=2.745, 95% CI 1.041-7.243). The variables related to LTFU or death in HIV patients were initial CD4 (HR=0.784, 95% CI 0.652-0.943), APRI≥0.5 (HR=4.647, 95% CI 1.331-16.227), OIs (HR=4.910, 95% CI 2.352-10.247) and age (HR=1.336, 95% CI 1.004-1.778). HBV coinfection was not associated with increased LTFU or overall mortality in HIV patients (p>0.05). Conclusion: With good ART treatment and adherence rate, the clinical indicators in HIV/HBV co-infected patients. the incidence in these baseline among were examined using the test for non-parametrically distributed variables, and t test for normally distributed variables. Categorial variables compared via the χ 2 test. Analysis of covariance (ANCOVA) for repeated measures was used to analyze the changes of biochemical indicators in HIV/HBV co-infected patients. Multiple logistic regressions (variable selection implemented via stepwise approach) was used to analyze the influencing factors of the abnormal incidence of clinical indicators and Cox model was used to analyze the influencing factors of LTFU and death in HIV patients. Statistical analyses were conducted with SAS 9.4 software.


Introduction
Liver diseases that are caused by the hepatitis B virus (HBV), including cirrhosis and hepatocellular carcinoma (HCC) [1,2], have become increasingly important in patients infected with HIV as their life expectancy is getting longer with successful ART [3][4][5]. Worldwide, approximately 37 million people are infected with HIV and about 5-20% are co-infected with HBV [6,7]. By the end of Sep. 2018, a total of 849,602 HIV patients were alive in China [8], 67.3% of whom had received antiretroviral therapy (ART) treatment [9], and the prevalence of HIV and HBV coinfection was 8.7-14% [4,10]. HIV and HBV shared transmission routes, including intravenous drug abuse, blood transfusion or blood products, sexual and mother-tochild transmission, whereas the immunodeficiency also caused by HIV enhances the likelihood of HBV persistence [1,2,7]. All these reasons lead to a gradual increase in the incidence of HIV/HBV co-infected patients.
Another reason may be related to hepatitis B vaccine which is an excellent intervention for preventing HBV infection [1,2,11]. At present, the hepatitis B vaccine vaccination rates in children under 5 years old remains above 90% in China, however, this rates in adults was fewer than 10% [11][12][13][14]. The rate of effective vaccination against HBV was only 7.7% in HIV-infected patients in west China [15].
In addition, HBV vaccine response rates is only 20 to 70% in HIV infected adults [16]. So vaccination of HBV vaccine in HIV high-risk population may help to reduce the incidence of HIV/HBV co-infection.
Among HIV patients receiving ART treatment, HBV status shows no effect on virological failure (OR=0.93, 95% CI 0.80-1.10) and CD4 counts (OR=0.93, 95% CI 0.82-1.05) [4,[17][18][19][20]. At present, more and more attention is being paid to mortality and hepatotoxicity in ART treatment for HIV/HBV co-infected patients. It is estimated that approximately 10-15% of mortalities in HIV patients globally are due to liver diseases, and about 25% of the liver deaths could be attributed to an HBV infection [1,21,22]. While the majority of those infected with HBV will naturally clear the infection, the risk of chronic HBV infection is estimated three-to six-fold higher in HIV patients than in those who are not infected with HIV [23][24][25]. ART induced liver toxicity is an additional concern in such patients. The incidence of hepatotoxicity in HIV patients taking ART is approximately 4.5-11% [18,26,27].
CD4 T-cell counts prior to ART initiation, compliance to HIV therapy, economic level, regional differences and other related factors, can affect the therapeutic effect of HIV or HIV/HBV patients [7,28,[38][39][40]. Especially in China, the economic level, health resource allocation, urban and rural factors among regions were greatly differentiated [15,41]. For example, the co-infection rate of HIV/HBV in west China was 14.4%, and it is 47% among men who have sex with men (MSM) population in some areas of China [7,15,42,43]. While the impact of ART on clinical outcomes in HIV/HBV infected patients has been studied, little is known about its comparative performance or efficacy in conventional clinical settings in China [7]. By the end of Sep. 2018, the total mortality rate of the 19,567 HIV patients who had received ART treatment in Beijing was 0.92% (180 cases), the total lost to follow-up (LTFU) rate was 1.6% (304 cases), and about 4.3% of all HIV patients are co-infected with HBV.

Study population
Data was retrospectively collected from the Beijing Center for Diseases Prevention Aspartate aminotransferase-to-Platelet Ratio Index (APRI) was categorized using validated cutoffs [28] : significant fibrosis (APRI≥0.5 ) and no significant fibrosis (APRI<0.5).ART adherence was reflected by the question of "Number of missed dose in recent 7 days", missed dose≥1 time (during follow-up, the HIV patient had at least one missed dose).

Follow-up
This study evaluated data from HIV patients and HIV/HBV co-infected patients

Statistical analysis
Mean ± SD or median and interquartile ranges (Q) were used to describe numerical values, and percentage (%) was employed to describe count data. Differences in baseline indicators among the two groups were examined using the Mann-Whitney U  Table 1. Median time interval (from diagnosis to anti-retro viral treatment), ALT and AST for HIV/HBV co-infected patients were statistically higher than those of HIV mono-infected patients. Baseline CD4 count, CD4/CD8 ratio, platelet and hemoglobin for HIV/HBV co-infected patients were statistically lower than those of of HIV mono-infected patients.
As shown in Table 1, the incidence of OIs (9.3% vs 4.5%) and APRI≥0.5 (7.9% vs 1.8%) in HIV/HBV co-infected patients was higher than that in HIV mono-infected patients (P<0.05, Table 1). The mean follow-up time for all patients was 34.5 months (SD: 24.2 months) per person in our study. The trajectory of key biochemical indicators plotted alongside ART treatment time is shown in Figure 1. The effect of baseline data on indicator changes during ART was adjusted by covariance analysis for repeated measures.
Although, CD4 count levels were still statistically different between HIV monoinfected and HIV/HBV co-infected patients after 42 months of ART treatment (P<0.05), no statistical difference was found between the two groups in the ART treatment effect after adjusting the influence of the baseline CD4 count (P=0.3226, Figure 1A); In the course of treatment, the TG and TC means in HIV/HBV co-infected patients were lower than that in HIV mono-infected patients (P g <0.05, Figure 1F-G).
Besides, with the increase of ART treatment time, ALT in HIV/HBV patients was different from that in HIV patients at 3 months. And 3 months later, ALT gradually decreased and there was no difference with HIV mono-infected patients (P g*t <0.001, Figure 1D). As for other indicators, AST, GLU, Cr, CD4/CD8 ratio, platelet and hemoglobin, the interaction between group and time effects was not significant (P>0.05).    [4,21,28]. We observed that HIV diagnosis to ART initiation interval, the initial incidence of OIs (9.3%) and baseline clinical characteristics ALT, AST for HIV/HBV co-infected patients were higher than those of HIV mono-infected patients. Moreover, baseline CD4 count, CD4/CD8 ratio, platelet and hemoglobin were lower in these patients. And, HBV coinfection affects the incidence of APRI≥0.5, but does not affect the incidence of CD4<200 cells/μL, CD4/CD8 ratio<0.30 and VL>400 copies/mL in HIV patients after the ART (Table 2).
After adjusting the effect of baseline CD4 count on indicator changes during ART, we found that CD4 count was still lower in HIV/HBV co-infected patients after ART treatment, but this difference was mainly affected by baseline CD4 count ( Figure   1A) and the prevalence of CD4<200 cells/μL was also not higher than that in HIV mono-infected patients (Table 2), which further illustrated HBV co-infection does not affect the therapeutic effect of CD4 in HIV patients [1,4]. Further investigation demonstrated initial low CD4 count, duration of ART, initial OIs and WHO stage increased the prevalence of CD4<200 cells/μL in HIV patients. Besides, initial CD4 count affects the incidence of VL>400 copies/mL and LTFU or death (HR=0.784) in HIV patients ( Table 2,3). Therefore, in order to better improve the ART effect of HIV/HBV patients, we should pay attention to the impact of these baseline CD4 [44,45].
The CD4/CD8 ratio could be used by clinicians to identity patients at risk of non-AIDS-related events and higher ratio may reflect a more " normal " immune phenotype conferring enhanced prognosis and predict posttreatment control [46][47][48][49]. In this study, although the baseline CD4/CD8 ratio in HIV/HBV patients was lower than that in HIV mono-infected patients, there was not significant difference between the two groups patients after ART treatment. Also we found initial low CD4 count, initial low CD4/CD8 ratio, shorter duration of ART and second-line ART increased the prevalence of CD4/CD8 ratio <0.30 in HIV/HBV patients. Evidence have shown that initial CD4 count can affect the ratio in HIV patients [46,[49][50][51].
Many studies have indicated that HIV/HBV coinfection may play a direct role in HCC, cirrhosis or liver fibrogenesis [1,32,52,53], and no clear reduction in end-stage liver disease (ESLD) risk was observed over 15 years in HIV patients, even after ART treatment [54]. We can see that baseline ALT, AST and the rate of APRI≥0.5 were higher in HIV/HBV co-infected patients (Table1). During ART treatment, the ALT of HIV/HBV patients increased significantly between 0 and 3 months. And 3 months later, ALT gradually decreased and there was no difference with HIV mono-infected patients ( Figure 1B). Besides, the rate of APRI≥0.5 in HIV/HBV co-infected patients was also higher, which mainly affected by initial APRI and HBV co-infection [55].In our study we also found that initital APRI≥0.5 (HR=4.6471, 95% CI 1.331-16.227) predicted for LTFU or death in HIV patients. Some studies have shown that higher AST appear to be important mortality risk factors in HIV/HCV-coinfection [56], both AST ,ALT and APRI to predict liverrelated mortality, either alone or as components of indices of liver fibrosis [28,53,[56][57][58]. Besides, several studies have shown a significant correlation between APRI scores and HIV viremia levels [28,59,60].
Therefore, the changes of AST,ALT and APRI in HIV/HBV patients after ART treatment need to be noticed, early detection of possible liver injury in patients and adjustment of ART regimens.
Currently, we can observe an increasing number of studies focused on the non-AIDS complications in HIV patients [61][62], analyzing changes in biochemical indicators (e.g., cholesterol) in patients undergoing ART treatment [64,65].We found that in the course of ART treatment, the TG and TC means in HIV/HBV co-infected patients were lower than that in HIV mono-infected patients ( Figure 1F-G),but the ratio of TC>5.69 mmol/L was no difference between them. HIV patients with ART are at increasing cardiovascular disease (CVD) risk and stroke, thus, the possibility to classify patients for lipid lowering treatment may be a useful tool for clinical management [66][67][68][69].But other studies showed that HIV patients with high serum total cholesterol have lower HIV RNA load and better CD4 T cell [65,70]. Therefore, there is need for prospective cohort or case control studies to determine the relationship between total cholesterol and ART outcomes in HIV patients. ART also confer an increased risk of early mortality [78].Besides, the impact of age on HIV patients has always been a topic.An exaplanation for this trend may be that older patients have poorer immunological responses than their younger counterparts [79]. Furthermore, successful ART has led to a growing number of older HIV-1-infected individuals who face both age-related and HIV-1-related inflammation, which may synergistically promote physical decline [80,81].We also found that HBV co-infection had no effect on HIV patients' loss of follow-up or death [4,32,56].
Limitations of this study were also considered. Indeed, due to individual differences, it is impossible to collect timely drug replacement information of patients, so we analyzed the ART treatment of HIV/HBV patients as a whole. Additionally, laboratory examination of HIV patients could not guarantee that they would be checked during each follow-up. Finally, there are few data on mortality in HIV/HBV patients in this study. Further long-term study are needed about the ART effect of HIV/HBV disease progression.
In this cohort analysis of HIV/HBV co-infected patients on ART outcomes, we found that clinical indicators such as CD4 count, VL, CD4/CD8 ratio, AST, the prevalence of OIs and APRI≥0.5 were improved significantly over time on ART treatment. However, the abnormal rate of AST and APRI was higher in HIV/HBV patients, and CD4, TC and TG were lower than those in HIV patients. HBV co-infection affect the prevalence of

Consent for publication
Not applicable

Availability of data and materials
The data that support the findings of this study are available at Beijing Center for Diseases Prevention and Control (Beijing CDC) and may be obtained from the authors upon reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This work was supported by grant from National Natural Science Foundation of China (project no. 71874100) and Beijing Municipal Science &Technology Commission (project no. D171100006717002).

Authors' contributions
All authors take responsibility for the structure of this paper. PW, BG and JX conducted the literature review and data analysis. PW drafted the paper. All authors contributed to the study's conception and design, interpretation of the data, and critical revisions to the paper. All authors have approved the final version for submission. Additional file 1.docx