Slow CD4+ T-Cell Recovery in Human Immunodeficiency Virus/Hepatitis B Virus-Coinfected Patients Initiating Truvada-Based Combination Antiretroviral Therapy in Botswana

Background. Hepatitis B virus (HBV) and human immunodeficiency virus (HIV) coinfection has emerged as an important cause of morbidity and mortality. We determined the response to Truvada-based first-line combination antiretroviral therapy (cART) in HIV/HBV-coinfected verus HIV-monoinfected patients in Botswana. Methods. Hepatitis B virus surface antigen (HBsAg), HBV e antigen (HBeAg), and HBV deoxyribonucleic acid (DNA) load were determined from baseline and follow-up visits in a longitudinal cART cohort of Truvada-based regimen. We assessed predictors of HBV serostatus and viral suppression (undetectable HBV DNA) using logistic regression techniques. Results. Of 300 participants, 28 were HBsAg positive, giving an HIV/HBV prevalence of 9.3% (95% confidence interval [CI], 6.3–13.2), and 5 of these, 17.9% (95% CI, 6.1–36.9), were HBeAg positive. There was a reduced CD4+ T-cell gain in HIV/HBV-coinfected compared with HIV-monoinfected patients. Hepatitis B virus surface antigen and HBeAg loss was 38% and 60%, respectively, at 24 months post-cART initiation. The HBV DNA suppression rates increased with time on cART from 54% to 75% in 6 and 24 months, respectively. Conclusions. Human immunodeficiency virus/HBV coinfection negatively affected immunologic recovery compared with HIV-1C monoinfection. Hepatitis B virus screening before cART initiation could help improve HBV/HIV treatment outcomes and help determine treatment options when there is a need to switch regimens.

Hepatitis B virus (HBV) and human immunodeficiency virus (HIV) coinfection represents a considerable health burden worldwide. As combination antiretroviral therapy (cART) has greatly improved survival in HIV-infected people, HBV has emerged as a major cause of morbidity and mortality in this group [1][2][3]. It is estimated that 5%-20% of the 35 million people living with HIV are also infected with HBV [4]. Sub-Saharan Africa has the highest burden of HIV/HBV coinfection [4]. In Botswana, HIV/HBV coinfection prevalence is 5.3%-10.6% [5,6]. Human immunodeficiency virus/HBV coinfection worsens disease outcome more than having either infection alone. Human immunodeficiency virus has been shown to change the natural course of chronic HBV by increasing the likelihood of HBV infections acquired during adulthood to progress to chronic HBV infection [7], increasing rates of hepatitis e antigen positivity [8] and higher levels of HBV DNA [9] but lower alanine aminotransferase (ALT) levels and rapid liver disease progression [2]. Higher mortality has also been reported in HBV/HIV coinfection compared with HBV-monoinfected individuals [10,11]. On the other hand, HBV has been reported to lead to higher baseline HIV viral load, lower CD4 + T cells, increased occurrences of advanced disease, lower body mass index (BMI), and increased mortality in HIV/ HBV-coinfected participants compared with HIV-monoinfected individuals [12][13][14][15]. However, these data have not been replicated in other studies [16,17].
The first-line HIV cART regimens may include drugs such as lamivudine, tenofovir, and emtricitabine that act against both HIV and HBV [18]. Screening for HBV before highly active antiretroviral therapy (HAART) initiation is standard of care in developed countries but not in resource-limited countries due to cost [15]. Hence, only HIV response is usually monitored in such settings. Most studies done on HBV/HIV coinfection are from low HBV-endemic areas, and these data may not be generalizable because the times of acquisition of HBV infection is in adulthood, whereas that of the high-endemicity areas is usually in early childhood [19]. Moreover, data from East Asia may not be applicable to Africa because of the different circulating genotypes and high HBeAg positivity in East Asia [19]. There are some studies that have been done to show the response of HBV to tenofovir, but most of them are not in Africa [17,20,21]. Most of the studies done in Africa were on response to lamivudine-containing regimens [1,15,22,23]; however, starting from 2010, the World Health Organization recommended that the first-line regimen for HIV should contain tenofovir [24]. There is increasing access to tenofovir; thus, data on HBV response to tenofovir is important, especially in high HBV-endemic and resource-limited countries. There are conflicting data on the effects of HBV on HIV response to cART [1,10,11,22,23]. In this study, we determined HBV response to tenofovir containing first-line cART in HIV-infected individuals initiating cART by determining HBV surface antigen (HBsAg) loss, HBV e antigen (HBeAg) loss, and HBV deoxyribonucleic acid (DNA) levels at different time points during treatment. We also determined the effects of HBV coinfection on HIV response to cART.

Study Participants
This was a retrospective longitudinal study. The research used archived plasma samples from HIV-infected adults who were initiating HAART. The samples were from a previous cohort-the Botswana National Evaluation Models of HIV Care (Bomolemo study): A Study Evaluating the Efficacy and Tolerability of Tenofovir and Emtricitabine (Truvada) as the Nucleoside Reverse Transcriptase Inhibitor (NRTI) backbone as first-line HAART for adults in Botswana-conducted by the Botswana Harvard AIDS Institute Partnership. This study enrolled 309 HIV-positive, treatment-naive adults who are 18 years or older and had CD4 + T-cell counts less than or equal to 250 or an acquired immune deficiency syndromedefining illness. The participants signed an informed consent for enrollment into the study. The study was approved by the University of Botswana Institute Review Board and Human Research Development Committee at the Botswana Ministry of Health.

Hepatitis B Virus Serologic Screening
The study screened 300 samples, which were available in storage (sample size was calculated based on the latest HBV/HIV prevalence in Botswana reported in literature [5,25]). All samples were screened for HBsAg using the Murex HBsAg version 3 kit (Murex Biotech, Dartford, UK) according to the manufacturer's instructions. Samples positive for HBsAg were then screened for HBeAg using the Monolisa HBeAg-Ab PLUS kit (Bio-Rad, Hercules, CA). Both HBsAg and HBeAg losses were evaluated at 6, 12, 18, and 24 months. Hepatitis B core antibody (HBcAb) and hepatitis surface antibody (HBsAb) were evaluated at baseline using Monolisa Anti-HBC PLUS (Bio-Rad, Marnes-la-Coquette, Paris, France) and Monolisa Anti-HBS PLUS (Bio-Rad, Marnes-la-Coquette, Paris, France) enzymelinked immunosorbent assay kits according to the manufacturer's instructions, respectively.

Hepatitis B Virus Genotyping
The HBV genotypes were determined at baseline as previously described [26]. In brief, a 415-base pair fragment of the s gene was amplified in a seminested polymerase chain reaction using HBV 381 primer (5′-TGCGGCGTTTTATCATCTTCCT-3′; nucleotide [nt] 381-402) and HBV 840 primer (5′-GTTTAAA TGTATACCCAAAGAC-3′; nt 840-861) for the first round [27]. The HBV 381 and HBV 801 primers (5′-CAGCGGCAT AAAGGGACTCAAG-3′; nt 801-822) were used for second round [27]. The HBV genotypes and resistance mutations were determined using online databases (Stanford and Geno2-Pheno) [28,29] and confirmed by phylogenetic analysis. The evolutionary history was inferred by using the maximum likelihood method based on the general time-reversible model. Initial trees for the heuristic search were obtained automatically by applying neighbor-joining and BioNJ algorithms to a matrix of pairwise distances estimated using the maximum composite likelihood approach and then selecting the topology with superior log likelihood value. A discrete gamma distribution was used to model evolutionary rate differences among sites. The rate variation model allowed for some sites to be evolutionarily invariable. Evolutionary analyses were conducted in MEGA6.06 [30]. Sequences have previously been submitted to National Center for Biotechnology Information GenBank under accession numbers KR139680 to KR139749 [26].

Quantification of Hepatitis B Virus Deoxyribonucleic Acid
Hepatitis B virus DNA levels were determined on samples that were HBsAg positive at baseline using COBAS AmpliPrep/ COBAS TaqMan HBV Test, version 2.0 (Roche Diagnostics, Mannheim, Germany) with a limit of detection of 20 IU/mL. The HBV DNA levels were also determined at 6, 12, and 24 months posttreatment initiation. Hepatitis B virus DNA suppression was defined as undetectable HBV DNA. Virological breakthrough was defined as an increase of >1 log 10 IU/mL in HBV DNA level from nadir in 2 consecutive measurements [31].

Data Analysis
The relationship between HBV status and HIV ribonucleic acid (RNA) suppression (RNA <400 copies/mL), CD4 + T-cell count, liver enzymes (ALT and aspartate aminotransferase [AST]), BMI, and mortality were assessed using Wilcoxon rank-sum test and logistic techniques.
We could not classify this as virological breakthrough because we could not ascertain whether the true value of HBV DNA level was ≥10 IU/mL because it was below limit of detection of the assay.

Hepatitis B Virus Genotypes
The HBV genotypes were 24 (85.7%) genotype A and 4 (14.3%) genotype D (Figure 3). At baseline, no HBV resistance mutations were found. The median baseline CD4 + T-cell count in participants with HBV genotype A was 293 cells/mL (IQR, 216-397), whereas median baseline CD4 + T-cell count in participants with HBV genotype D was 274 cells/mL (IQR, 198-309). The results suggested no significant difference in CD4 T cells between the 2 genotype groups (P = .1366). There was an overall significant difference in CD4 + T-cell increase from baseline after 24 months between HBV genotypes A and D (P = .0019). Subjects with HBV genotype A had an overall average CD4 + T-cell increase of 144 cells/mL (95% CI, 120-169), whereas HBV genotype D subjects had an average CD4 + T-cell increase of 55 cells/mL (95% CI, 7-103  Figure 4). However, the difference in HBV viral load between   HBV genotypes after 24 months did not reach statistical significance (Fisher's exact P = .437).

DISCUSSION
In this study, we report for the first time the impact of Truvadabased cART on HIV/HBV-coinfected patients compared with HIV-1-monoinfected patients in Botswana. Overall, both coinfected and HIV-1-monoinfected patients responded well to the Truvada-based cART, with HIV-1 infection markers  CD4 + T-cell counts and HIV-1 viral load showing positive response. Likewise, HBV infection markers HBsAg and HBeAg were lost in some of the participants. Furthermore, HBV viral load was also suppressed in most participants. However, the HIV/HBV-coinfected patients had a slower CD4 + T-cell gain compared with the HIV-1-monoinfected patients.
The HIV/HBV prevalence is comparable to previous reports within the region, which range from 4.2%-22.9% in Zimbabwe and South Africa [6,[32][33][34]. We also found similar rates of HBeAg, HBcAb, and HBsAb as reported previously in Botswana [6].
The HBsAg loss documented in this study (38%) was higher than what has been reported in other studies [35][36][37]. The high and rapid HBsAg loss might be partly due to genotypes A and D, which have been associated with a rapid initial HBsAg reduction that correlates with HBsAg loss [38]. The study conducted in South Africa and Zambia also had a higher HBsAg loss (18%) within 12 months [35] compared to studies such as those in the Ivory Coast, where HBsAg loss was as low as 6.2% after a median of 12.3 months on treatment [39], but it was lower than this cohort. The difference might be the differences in genotypes because 22% of the Hamers study were genotype E [35].
However, the Hamers study only checked for HBsAg loss at 12 months and not at 6 months [35]. Furthermore, Zoutendijk et al [37] noted that most of HBsAg loss happens during the first year [40], which might explain the rapid HBsAg loss in our study. On the other hand, the HBV DNA suppression is comparable to other studies [35,41]. There are other studies that have also demonstrated that HBV virologic suppression increased with time on treatment [42]. Baseline CD4 + T cells were significantly associated with HBV DNA suppression at 12 months but not at other visits. This concurs with other studies such as Kim et al [43] who reported baseline CD4 + T cells as a predictor of HBV suppression in response to tenofovir [44].
The pretreatment CD4 + T-cell counts in this study were similar between the 2 groups as reported by Kim et al [43] in Kenya [10]. This was also similar to another study that included Botswana and South African pregnant women, and the baseline CD4 + T cells for Batswana were similar between the HIV/HBVcoinfected and HIV-monoinfected group, whereas that of South Africans were lower in the HIV/HBV-coinfected group [45]. Other studies in South Africa and Tanzania reported marginally lower baseline CD4 + T cells [17,35]. Some studies found similar CD4 + T-cell response in both groups [1,15,23,35]. However, we documented a lower CD4 + T-cell response in HIV/HBVcoinfected participants compared with HIV-monoinfected participants, similar to the results found in the large cohort from Tanzania [17]. This large cohort from Tanzania also reported a marginally significant recovery at 12 months; however, in our cohort, there were no significant differences in immunologic recovery between HIV/HBV-coinfected and HIV-monoinfected participants at 12 months [17]. The differences might be due to the smaller number of the participants in our cohort. Furthermore, another study in Southern Africa also reported no difference in CD4 + T-cell recovery at 12 months after cART initiation [35]. The lower CD4 + T-cell response in HIV/HBV-coinfected   could be due to destruction of CD4 + T cells by HBV-mediated T-cell activation [46]. Comparable hepatotoxicity between HIV/HBV-coinfected and HIV-monoinfected participants was observed, as reported by others [16,[47][48][49]. However, some studies in South Africa, Ghana, Kenya,Tanzania, and Cambodia have shown increased hepatotoxicity in HIV/HBV-coinfected participants [1,11,17,22,23].
Pretreatment HIV-1 RNA was similar in both HIV/HBVcoinfected and HIV-monoinfected participants, as reported in other studies [1,15,22]. On the other hand, higher baseline HIV RNA in HIV/HBV-coinfected participants has been documented [13]. In this study, similar to studies conducted in South Africa and Kenya, there was no difference in HIV viral load suppression between the 2 groups [1,15,22]. This is contrary to a study conducted in Taiwan, which reported a lower HIV viral load suppression in HIV/HBV-coinfected participants compared with the HIV-monoinfected group [50].
The HBV genotypes found in this cohort are consistent with what have been reported in the country [26] and within the region [51]. There was no significant difference in HBV DNA level at all time points between genotype A and D. The lack of association between HBV genotype and HBV DNA load has been reported before [12,52], but in our cohort it might be due to small numbers. However, some studies have recorded differences between genotypes, and genotype D was found to have a higher HBV viral load [53]. Genotype D participants had lower CD4 + T-cell responses than participants with genotype A. Genotype D has been associated with a worse outcome in some studies [54,55], which might explain the worse immunosuppression in these individuals.

LIMITATIONS
The limitations of this study include small sample size. The HIV/HBV prevalence might be underestimated, because it does not include occult HBV. There is a need in future to generate data in a larger sample size that can be extrapolated to the general population. Furthermore, the cohort was predominately female, hence data may not be generalizable to men. Another limitation of this study is that HBcAb immunoglobulin M screening, a test for hepatitis acute infection, was not performed; however, most of the infections in sub-Saharan Africa are acquired during childhood, and this was a cohort of adults. Hence, we could assume that most of the infections were chronic infections. The clinical significance of unsuppressed HBV DNA need to be studied further in this setting.

CONCLUSIONS
In conclusion, HIV/HBV coinfection might have a negative effect on the immunologic response of HIV-1 to cART; hence, screening and monitoring of HIV/HBV-coinfected participants might be important in this setting. Most participants suppressed HBV DNA, and the HBV serologic response was higher than in most studies. This supports the recommendation of Truvada-based cART in HIV/HBV coinfection. However, because approximately 25% of the participants did not completely suppress HBV DNA after the 2-year follow up, it is crucial to monitor HBV response in coinfected patients on Truvadabased cART so that in case of treatment switch for HIV virologic failure, the patients are not put on a regimen that is less effective against HBV and could result in the rebound of HBV viremia.