Prevalence of hepatitis B and C viruses in HIV-positive patients in China: a cross-sectional study

Introduction Liver disease related to hepatitis B (HBV) and hepatitis C (HCV) may temper the success of antiretroviral therapy (ART) in China. Limited data exist on their prevalence in HIV-positive Chinese. A multi-centre, cross-sectional study was carried out to determine the prevalence and disease characteristics of HBV and HCV co-infection in HIV-positive patients across 12 provinces. Methods HIV-positive ART-naïve patients were recruited from two parent cohorts established during November 2008–January 2010 and August 2012–September 2014. Hepatitis B surface antigen (HBsAg), hepatitis B e antigen and HCV antibody (anti-HCV) status were retrieved from parent databases at the visit prior to ART initiation. HBV DNA was then determined in HBsAg+ patients. HCV RNA was quantified in anti-HCV+ patients. Aspartate aminotransferase-to-platelet ratio index (APRI) and the fibrosis-4 (FIB4) were calculated. Chi-square test, Kruskal–Wallis test and logistic regression were used for statistical analysis, as appropriate. Results Of 1944 HIV-positive patients, 186 (9.5%) were HIV–HBV co-infected and 161 (8.3%) were HIV–HCV co-infected. The highest HIV–HBV prevalence (14.5%) was in Eastern China while the highest HIV–HCV prevalence was in the Central region (28.2%). HIV–HBV patients had lower median CD4 + T cell count (205 cells/μL) than either HIV monoinfected (242 cells/μL, P=0.01) or HIV–HCV patients (274 cells/μL, P=0.001). Moderate-to-significant liver disease was present in >65% of the HIV–HCV, ~35% of the HIV–HBV and ~20% of the HIV monoinfected patients. Independent associations with moderate-to-significant liver disease based on APRI included HBV (Odds ratio, OR 2.37, P < 0.001), HCV (OR 9.64, P<0.001), CD4 count≤200 cells/μL (OR 2.55, P<0.001) and age ≥30 years (OR 1.80, P=0.001). Conclusions HBV and HCV prevalence is high in HIV-positive Chinese and differs by geographic region. HBV and HCV co-infection and HIV monoinfection are risks for moderate-to-significant liver disease. Only HIV–HBV is associated with greater HIV-related immunosuppression. Incorporating screening and management of hepatitis virus infections into Chinese HIV programmes is needed.


Introduction
In the era of antiretroviral therapy (ART), liver disease from hepatitis virus co-infection is a leading cause of morbidity and mortality in the HIV-positive population in North America and Europe [1,2]. Although hepatitis virusÁHIV co-infection has been extensively studied in developed countries [1,3Á5], it has not been well characterized in Asia and Africa where these viruses are highly endemic [6Á8]. There is concern that hepatitis virusÁrelated liver disease may threaten the success of ART programmes in developing countries; therefore, understanding the prevalence and disease characteristics of HBV and HCV co-infection with HIV in Asia and Africa is essential.
China has a heavy disease burden of both chronic viral hepatitis infections and HIV/AIDS. In the general Chinese population, the prevalence of hepatitis B surface antigen (HBsAg) in those aged 1Á59 years is 7.2% [9] while the estimated HCV prevalence is 1.0Á2.9% [6,7,10]. However, the prevalence of these viruses in the 780,000 people with HIV/ AIDS in China is not known. A National Free Antiretroviral Treatment Program (NFATP) was initiated in 2002. By 2013, 227,489 patients were receiving ART through the programme [11]. However, the HBV and HCV status were unknown in !50% of the patients starting ART between 2010 and 2011 in NFATP [12]. Previous studies have only been conducted at a single site or focused on certain populations, so they do not provide information on the profile of hepatitis virusÁ HIV co-infection across China [13Á15]. Limited multi-centre studies showed that HBsAg seropositivity ranged from 8.7 to 12.5% while seroprevalence of HCV antibody (anti-HCV) varied from 12.2 to 41.8% [12,16,17]. However, these studies did not discuss the HIV or hepatitis disease characteristics.
To determine the nationwide prevalence and disease characteristics of HBV and HCV co-infection in HIV patients, a multi-centre study was carried out using patients enrolled in various China AIDS Clinical Trials in 12 provinces.

Study design and participants
This was a cross-sectional study that included ART-naïve participants enrolled in one of the following parent studies: China AIDS Clinical Trial (CACT) 0810 or CACT1215 (ClinicalTrials. gov identifier: NCT00872417 and NCT01844297). For this study, only study entry data prior to ART were included. Participants in CACT0810 were enrolled between November 2008 and January 2010 (n0879). As previously described [18], inclusion criteria in the cohort were: age: 18Á65 years, CD4' T cell count B350 cells/mL and ART-naïve. Exclusion criteria included: acute HIV infection, currently active AIDS-defining illness, pregnancy or breastfeeding and active intravenous drug use (IDU). Participants in CACT1215 were enrolled between August 2012 and September 2014 (n01191). CACT1215 had the same inclusion criteria as CACT0810, except for the CD4' T cell count threshold of 500 cells/mL. The Institutional Review Board of Peking Union Medical College Hospital (PUMCH) approved the parent studies and each participant provided written informed consent. All the data and specimens used in the present study were retrieved from parent databases before ART initiation.
Study participants were from one of twelve provinces or municipalities across China, which were clustered into five geographic regions ( Figure 1): North, Beijing and Liaoning; East, Shanghai and Zhejiang; South, Guangdong, Guangxi and Fujian; Central, Henan and Hunan; West, Shaanxi, Sichuan and Yunnan. Demographic and clinical data before ART initiation, including age, sex, HIV transmission route, alanine transaminase (ALT), aspartate aminotransferase (AST), platelet count, total bilirubin (Tbil), CD4 ' T cell count, HIV RNA, HBsAg, hepatitis B e antigen (HBeAg) and anti-HCV serostatus were abstracted from the parent databases. HBV and HCV serology tests were performed using various commercialbased kits approved by the China Food and Drug Administration. The present study was approved by the Institutional Review Board of PUMCH.
In HBsAg' patients, HBV DNA was determined. HCV RNA was quantified in anti-HCV' patients. These values were determined prior to ART initiation. Participants were defined as having HIVÁHBV co-infection if they were HBsAg'anti-HCV ( or HBsAg' anti-HCV'HCV RNA (. Participants negative for HBsAg and positive for both anti-HCV and HCV RNA were defined as HIVÁHCV co-infected. Participants positive for HBsAg, anti-HCV and HCV RNA were grouped as triply infected. Participants were identified as HIV monoinfected when they were negative for both HBsAg and anti-HCV or HCV RNA.
Two non-invasive markers, AST-to-platelet ratio index (APRI) [19] and the fibrosis-4 score (FIB4) [20], were used to evaluate the liver fibrosis of study participants. APRI was calculated according to Wai et al. [19]: (actual AST value divided by its upper normal limit considered as 40 U/L)/platelet counts (10 9 /L))100. FIB4 score was calculated according to Sterling et al. [20]: [age (years))AST (U/L)]/[platelets (10 9  . APRI and FIB4 scores were ranked into three classes in accordance with previous studies conducted in HIV-positive population: APRI class 1, 50.5; APRI class 2, from 0.51 to 1.5; APRI class 3, !1.5; FIB4 class 1, 51.45; FIB4 class 2, from 1.46 to 3.25; FIB4 class 3, !3.25 [22]. Comparisons among groups were made by non-parametric methods. Chi-square or Fisher's exact test was used for categorical variables and KruskalÁWallis test or MannÁWhitney U test was used for continuous variables. Estimated prevalence and 95% confidence interval (CI) were calculated by modified Wald method. Separate univariate binary logistic regression models were used to assess the odds ratio (OR) for elevated liver fibrosis scores with the following cutoffs: APRI ! 0.5 and FIB4 ! 1.45 (elevated scores). Parameters included in the models were age, sex, CD4' T cell count, HIV RNA levels and hepatitis virus co-infection status. Multivariable logistic regression models were then developed using covariates that were significant (PB0.05) in the univariate models. The analysis was performed among the entire cohort and then restricted to the HIV monoinfected group to identify predictors for increased liver fibrosis scores. Statistical analysis was performed using SPSS 22.0 (IBM Corporation, Armonk, New York, USA) and GraphPad Prism 6.0 (GraphPad Software, Inc. La Jolla, CA, USA). P value B0.05 was considered statistically significant.

Results
Demographics of participants and prevalence of HBV and HCV in HIV-positive patients A total of 2070 treatment-naive participants were eligible for this analysis, 74 (3.6%) of whom were excluded because of missing HBsAg and/or anti-HCV results. A further 48 (2.4%) anti-HCV' patients were excluded because specimens were not available for HCV RNA testing leaving 1948 included in this study. The excluded participants were similar in age, CD4 count and sex to the included participants (data not shown). Of the 1948 participants, 186 (9.5%) were HIVÁHBV coinfected, 161 (8.3%) were HIVÁHCV co-infected with a mean HCV RNA level of 6.3491.0 log 10 IU/mL and 4 (0.2%) had triple infection. Triply infected participants were not included in further analyses because of the small number of subjects.
Interestingly, the prevalence of HIVÁHBV and HIVÁHCV co-infection varied widely by region ( Figure 1 and Table 2). Participants in Eastern China had the highest prevalence of HIVÁHBV co-infection (14.5%), while the Central region had the lowest of 5.0%. In contrast, the Central region had the highest HIVÁHCV prevalence (28.2%) followed by the West (11.5%), the North (4.6%), the East (2.2%) and finally the Southern region (2.0%). In the Central region, 50.2% of participants were infected by blood transfusion, which was significantly higher than that in the other four regions ranging from 11.5% in the North to 1.5% in the South.

Liver disease characteristics
Hepatitis virus co-infected participants had higher median ALT and AST values than HIV monoinfected participants (Table 3).
Notably, the highest ALT and AST values were seen in the HIVÁHCV co-infected participants with Â10% having moderate or severe elevations. The HIVÁHCV co-infected participants also had the highest median APRI score (0.80), followed by HIVÁHBV co-infected (0.40, P B0.001) and HIV monoinfected participants (0.32, P B0.001) ( Table 3). Nearly threequarters of the HIVÁHCV co-infected participants had APRI scores !0.50, which is considered as moderate-to-significant hepatic fibrosis. This proportion was much higher than participants with HIVÁHBV co-infection (36.9%, P B 0.001) or HIV monoinfection (20.8%, PB0.001). Similar results were seen with FIB4 ( In the 186 HIVÁHBV co-infected participants, 57 (30.6%) were positive for HBeAg. Age, sex and distribution of HIV transmission routes did not differ between participants with and without HBeAg ( Table 5). HBeAg-positive patients trended towards lower CD4' T cell counts (median 188 cells/mL, IQR 106Á283 cells/mL) than HBeAg-negative participants (median 214 cells/mL, IQR 132Á312 cells/mL, P00.21). No difference was found in HIV RNA levels between the two subgroups. Plasma to measure HBV DNA was available in 43 (75.4%) HBeAgpositive and 105 (81.4%) HBeAg-negative participants. The median HBV DNA was 8.03 log 10 IU/mL in the HBeAg-positive participants, which was higher than in the HBeAg-negative participants (3.02 log 10 IU/mL, PB0.001). HBeAg-positive participants had higher median ALT than HBeAg-negative participants (36 U/L vs. 27 U/L, PB0.001), but other markers of liver disease did not differ between the two groups ( Table 5).

Discussion
In this large study of HIV-positive persons across China, HBV and HCV co-infection was found in 9.5 and 8.3% of subjects, respectively. Notably, the distribution of co-infection was not uniform across the country and was different between HBV and HCV. We found that the HIVÁHBV co-infected participants were more immunosuppressed than HIV monoinfected or HIVÁHCV co-infected participants as demonstrated by having the lowest CD4' T cell counts. A large proportion of all subjects had moderate-to-significant liver disease as determined by the serum markers APRI and FIB4 with the greatest proportion in HIVÁHCV co-infected participants ( !65%). CD4' T cell count B200 cells/mL was an independent risk factor for elevated liver fibrosis scores regardless of hepatitis status. Taken together, these data demonstrate that HIV monoinfected persons can have moderate-to-significant liver disease and that co-infection with these hepatitis viruses is common in HIV-positive Chinese and leads to important clinical consequences, which are currently unrecognized.
The prevalence of HIVÁHBV co-infection in our study is consistent with previous multi-centre studies conducted in China ranging from 8.7 to 12.5% [12,16,17]. In accord with a previous multinational study [3], HIVÁHBV co-infected participants had lower CD4' T cell counts than HIV monoinfected participants. In contrast, the retrospective study with data from the China NFATP demonstrated that CD4' T cell counts did not differ between HIVÁHBV co-infection and HIV monoinfection [12]. One possible reason is that the NFATP study had more participants with CD4' T cell counts below 200 cells/mL and had a lower overall median CD4' Tcell count making it more difficult to find a difference in CD4' T cell counts between groups. Of the 186 HIVÁHBV co-infected patients, 30.6% were positive for HBeAg, which is markedly lower than previous studies from the United States (59%) [23], Canada (54%) [4] and the multinational study (50%) [3]. This difference may be due to differences in age of HBV acquisition. In countries such as the United States and Canada with low endemicity, HBV acquisition primarily occurs in adulthood. In contrast, the major mode of HBV acquisition in China is mother-tochild [24]; thus, they have been infected with HBV for years and so are more likely to have HBeAg-negative disease [25]. The Chinese national surveys indicate that the HBeAg prevalence inversely correlates with age and that the HBeAg prevalence in that survey for the age of our participants is similar to our findings [9,24]. Regardless of HBeAg status, over one-third of the HIVÁHBV co-infected patients had moderate-to-significant liver disease, so diagnosing and treating HBV is imperative as HBV treatment can improve liver disease [26,27].
The anti-HCV seropositivity rate (14.0%) in our cohort is within the range of 12.2 to 41.8% as seen in previous Chinese multi-centre studies [12,16,17]. However, our results, for the first time, demonstrate the prevalence of HCV co-infection of 8.3% since we tested HCV RNA in anti-HCV positive patients. Identifying patients with HIVÁHCV co-infection is important since we found that over two-thirds had moderate-tosignificant liver disease as measured by either APRI or FIB4. By non-invasive measurements, the prevalence of fibrosis in previous studies of HIVÁHCV co-infected patients is around 30Á50% in other countries [20,28,29]. One possible explanation for the higher prevalence of moderate-to-significant liver disease in our cohort is that our participants had lower median CD4' T cell counts, which we found to be an independent risk  factor for elevated APRI and FIB-4. The NFATP report showed that HCV co-infection is associated with higher mortality in the two-year period after ART initiation compared with those infected by HIV alone [12]. Taken together with our study, HCV co-infection may have clinical consequences in HIV-positive individuals on ART supporting the need for testing and treatment for HCV in ART treatment programmes in China. Even if HCV treatment is not readily available across China, these patients could be prioritized for HIV treatment regardless of CD4 cell count because some studies have shown that liver disease improves with ART [30,31].
The other notable finding in our study is that about 20% of the HIV monoinfected individuals had moderate-tosignificant liver disease. This is consistent with other studies from the United States but higher than in an international cohort [32,33]. We also found that liver disease in this group was strongly associated with CD4' T cell count 5350 cells/ mL using APRI and FIB4, as was demonstrated in a US-based study [32,33]. These data suggest that immunosuppression from HIV affects liver disease and supports earlier treatment of HIV disease. It is not clear why HIV monoinfected individuals have significant liver disease, but it may be related to HIV infection of hepatic stellate cells or hepatocytes, microbial translocation or an inflammatory state in the liver from HIV infection [34Á36]. Further work is also needed to determine whether ART improves liver disease in the HIV monoinfected individuals. This study has several limitations. First, patients were recruited from previous ART trials; thus, we cannot rule out that there were differences between patients who were included in parent trials and those not included. Second, active IDUs were not included in parent studies; thus, our results cannot be generalized to HIV-positive IDUs in China. However, since the current HIV epidemic is via sexual transmission [11], our results are timely. Third, inclusion criteria of parent studies may exclude patients with very high ALT or AST values, so we may have underestimated the number of co-infected people with advanced liver disease. Fourth, we could not collect data on other potential liver diseaseÁrelated factors such as alcohol use and metabolic syndrome. Regarding the difference of liver fibrosis in HIV monoinfected patients between this cohort and international studies, these hepatic comorbid factors should be taken into account. Finally, we did not have liver biopsy data and instead used APRI and FIB4 to evaluate liver fibrosis. Although these markers correlate well with liver disease in other countries, they have not been extensively validated in China. However, it is likely that they are representative of liver disease stage based on other studies.

Conclusions
This study is the first, to our knowledge, to characterize HBV and HCV co-infection in treatment-naïve HIV-positive patients in China. Our data demonstrate that there is a high prevalence of HBV and HCV co-infection and that co-infection is associated with important medical consequences. As ART scale up increases through China's treatment programme, liver disease from hepatitis virus co-infection and even from HIV monoinfection will be an emerging problem. Thus, incorporation of screening and effective treatment of hepatitis virus infection into general HIV management in China is imperative.