1. Introduction

An estimated 12 million people are seropositive for hepatitis delta virus (HDV) worldwide, accounting for a global prevalence of 0.16% [1]. Among hepatitis B surface antigen (HBsAg)-positive people, the anti-HDV prevalence ranges between 3.0% in Europe and 6.0% in Africa, with an overall estimate of 4.5% [1]. In Germany, an incidence of 0.07 per 100,000 in the general population was reported. Around one third of patients with newly diagnosed HDV are not of German descent but originate mostly from Eastern Europe [2]. People from HDV-endemic areas, people who inject drugs, hemodialysis recipients, men who have sex with men, commercial sex workers, and individuals infected with hepatitis C virus or human immunodeficiency virus carry an increased risk of infection with HDV [1]. Despite its relatively low prevalence in HBsAg-positive patients, chronic hepatitis D (CHD) accounts for 18% of cases with cirrhosis and 20% of cases with hepatocellular carcinoma (HCC) emphasizing the aggressive course of the disease [1]. Patients with CHD tend to have more severe liver disease, more rapid progression to cirrhosis, and an increased risk for hepatic decompensation and death compared to patients with hepatitis B virus (HBV) monoinfection [3]. The term compensated advanced chronic liver disease (cACLD) was recently introduced by the Baveno VI consensus report, taking into account the difficulties of differentiating between severe fibrosis and cirrhosis by means of clinical evaluation or non-invasive testing [4].

HDV is a small, defective ribonucleic acid (RNA)-virus which can propagate only with coexisting HBV since it depends on HBsAg for its assembly [3]. HBsAg interacts with the sodium taurocholate cotransporting polypeptide (NTCP), a sodium-dependent uptake transporter localized on the basolateral membrane of hepatocytes responsible for uptake of bile acids, which facilitates the entry of HDV into hepatocytes [5]. Treatment options for CHD are limited and pegylated interferon alfa (PEG-IFNα) used to be the only treatment option, other agents such as nucleos(t)ide-analogue (NA) are ineffective despite its efficacy in chronic hepatitis B [6–8].

Bulevirtide (BLV), a large lipopeptide comprising 47 amino acids, is a first-in-class inhibitor of NTCP [5, 9]. It mimics a region of the pre-S1 HBsAg and inhibits viral entry by inhibitory competition [9]. It is fully approved for the treatment of CHD in HDV-RNA positive adult patients with compensated liver disease at a daily dose of 2 mg by the European Medicines Agency (EMA) after conditional approval in July 2020 based on the results from clinical trials [10–13]. According to the summary of product characteristics, treatment should be continued as long as it is associated with a clinical benefit as the optimal treatment duration is unknown [14]. Discontinuation should be considered in case of sustained HBsAg seroconversion or loss of virological and/or biochemical responses [14]. Randomized controlled trials investigating BLV reported promising results regarding its efficacy and safety, however, for limited treatment durations of 24 or 48 weeks [10–12]. Data on long-term (>48 weeks) administration of BLV are sparse.

The objective of this study was to evaluate the treatment course of adult patients with advanced chronic liver disease (ACLD) who received treatment with BLV for at least one year in a real-world setting.

2. Materials and Methods

3. Results

3.2. Response to Bulevirtid Treatment

The median duration of BLV therapy at the last follow-up (FU) was 26 months (IQR 17–27). Treatment duration was ≥12 months in 14 (100%) patients, ≥18 months in 10 (71%), and ≥24 months in 8 (57%). One patient had received BLV for 33 months. After the first 12 months of treatment, HDV-RNA decreased from log₁₀ 5.58 IU/ml to log₁₀ 3.01 IU/ml (SD 1.32) and ranged between log₁₀ 2.19 IU/ml and log₁₀ 3.19 IU/ml, thereafter (Figure 1). Changes in ALT levels during treatment are shown in Figure 2, and changes in VR, BR, and CR are shown in Table 2. After 12 months of treatment, there was only one (7%) patient with NR who then experienced VR after 15 months, and one (7%) patient with PR who had not yet reached the next FU. Two (25%) patients showed an increase in HDV-RNA >1 log₁₀ IU/ml compared to previous viral load after 24 months of treatment but returned to previous HDV-RNA levels within the next FU. No effects on HBV-DNA and HBsAg were observed during the entire treatment period.

Table 2. Virological (VR), biochemical (BR), and combined response (CR) depending on time after start of bulevirtide; T: treatment month; LLQ: lower limit of quantification.

	T12	T15	T18	T21	T24	T27	T30	T33
VR (%)	86	100	90	100	75	100	100	100
HDV-RNA < LLQ (%)	43	50	50	57	63	57	0	0
BR (%)	43	60	60	86	75	86	50	100
CR (%)	43	60	60	86	75	86	50	100
Patients (n)	14	10	10	7 ∗	8	7	2	1

•  ^∗In one patient, who has been treated for 24 months, data for T21 were not available.

3.3. Adverse Events of Bulevirtid Treatment and Effect on Liver Function

Two patients reported mild and self-limiting adverse events (pruritus and abdominal discomfort), and treatment discontinuation was not necessary. Between baseline and last FU, the CP and MELD scores were maintained in 10 (71%) and 3 (21%) patients, improved in 2 (14%) and 9 (64%) patients, respectively, and worsened in 2 (14%) and 2 (14%) patients, respectively (Figure 3). The APRI at baseline was ≥1.09 in 11 (79%) patients, and the mean APRI significantly declined from 2.33 (SD 1.27) to 1.15 (SD 0.89) (p < 0.001) during treatment (Figure 4). Hepatic decompensation with ascites requiring large-volume paracentesis occurred in 1 (8%) patient with clinically significant portal hypertension after 24 months of treatment. At this time, the patient had a CR with HDV-RNA <LLQ, and there was no evidence of HCC or concomitant acute or chronic liver disease; therefore, BLV therapy was maintained. According to the BEA score, two (14%), eight (57%), and four (29%) patients were categorized as BEA-A, BEA-B, and BEA-C, respectively. The only decompensation event occurred in a patient classified as BEA-C during a cumulative FU of 319 months.

4. Discussion

In this retrospective monocentric study, BLV administration >12 months was associated with prolonged VR and BR as well as with preservation of liver function in patients with CHD and cACLD. Additionally, BLV was well tolerated and only a very limited number of side effects were observed.

CHD is a rare disease with a considerably lower prevalence than chronic hepatitis B and C [1, 18]. However, it is associated with a more aggressive course of liver disease and subsequently higher progression rates towards ACLD, portal hypertension, decompensation, and HCC [1, 3]. A preliminarily published German guideline on the treatment of CHD recommends evaluation of antiviral therapy in all patients with CHD with prioritization of patients with high inflammatory activity, advanced fibrosis, and compensated cirrhosis [19]. Until the conditional approval of BLV in July 2020, medical treatment options for CHD were mainly limited to PEG-IFNα-2a. Although PEG-IFNα-2a led to undetectable HDV-RNA in 24% and 33% after 48 and 96 weeks, respectively, half the patients experienced virological relapse after a median of 4.5 years [7, 8, 19], [20]. IFN treatment is also associated with frequent and occasionally severe adverse effects and has several contraindications. However, it was also associated with a lower risk of relevant clinical outcomes, such as hepatic decompensation and liver transplantation, compared to NA therapy alone [6]. Since persisting HDV viremia is a risk factor for cirrhosis and liver-related events, patients with CHD could benefit from viral suppression [21–23]. Hence, a decline in HDV-RNA of ≥2 log₁₀ was suggested as an indicator of an effective treatment [24].

BLV was the first drug approved for CHD treatment. It results in a notable decrease in HDV-RNA levels by engaging with NTCP, thereby, hindering the infection of previously uninfected hepatocytes. Because BLV does not interfere with the HDV life cycle within infected hepatocytes, relapse of HDV-RNA is anticipated as soon as BLV treatment is stopped [11]. Therefore, BLV should be continued as long as a clinical benefit is expected [14]. Furthermore, VR seems to increase with treatment duration which may be explained by a decreased burden of infected hepatocytes [12]. Randomized controlled trials reported VR of 54% after 24 weeks and 71% after 48 weeks with BLV at 2 mg per day [11, 12]. In several real-world cohort studies, VR ranged between 58% and 78% after 24–48 weeks of treatment [25–31]. Due to its fairly recent approval, experience with long-term administration of BLV is limited. To address this issue, we included patients with treatment duration >12 months only in our investigation. Thus, higher rates of VR and HDV-RNA negativity in our study may be best explained by the mode of action of BLV, but also because of a positive selection bias. Rates for BR and CR were higher in some, but not all previously published studies [25–31]. We observed corresponding rates for VB, which occurred in 9–14% [27, 28]. Both cases with VB showed significant reduction in HDV-RNA at the next FU visit. In these patients, bile acids were not available to evaluate treatment adherence. In line with previous studies, we did not observe changes in quantitative HBsAg levels during BLV treatment [27, 32].

Despite frequent reporting of adverse events (AEs) in randomized controlled trials (64–82%), AE grade 3 or 4 and serious AE were infrequent (10–11% and 0–4%, respectively) in BLV 2 mg treatment groups [11, 12]. Importantly, treatment discontinuation due to AE was not observed in our study. The most common AEs in the MYR301-trial, a phase 3 study, were headache, pruritus, fatigue, nausea, arthralgia, and injection-site events, e.g., erythema, pruritus, and swelling as well as changes in laboratory values [12]. In our analysis, BLV was well tolerated even during long-term administration in our study with no injection-site event, no serious AE, and no treatment discontinuation. The favorable safety profile of BLV could be attributed to its metabolism. As a peptide, BLV is likely degraded by proteases into amino acids without involvement of cytochrome enzymes [33]. Additionally, no active metabolites are expected.

Portal hypertension is a key determinant for the risk of hepatic decompensation in patients with ACLD and can be estimated by non-invasive testing, e.g., liver stiffness measurement (LSM) [34]. The majority of patients in our study had assumed or suggested cACLD with a corresponding high prevalence of portal hypertension. Accordingly, most patients had an APRI ≥1.09 at baseline, which had a positive predictive value of 85% for a hepatic venous pressure gradient >12 mmHg [35]. Whereas liver stiffness measurements were not regularly available during FU, we observed a decrease in APRI in all but one patient and a significant reduction in mean APRI. However, changes in APRI were mainly driven by changes in AST levels and not by an increase or decrease in platelet count. Therefore, it is questionable if improvements in APRI represent true improvements in portal hypertension or if this is due to a reduction in hepatic inflammation. The single patient with an increased APRI, who had only PR after 12 months of treatment, also experienced an increase in liver stiffness from 9 kPa to 14 kPa, which was confirmed on separate occasion. Degasperi et al., however, described similar findings with improvement in AST and unchanged platelet count after 48 weeks of BLV treatment in patients with ACLD and, additionally, a significant decrease in liver stiffness [30].

An important finding of our study is that liver function, represented by Child Pugh and MELD scores, was stable or improved in the majority of patients during long-term treatment with BLV. Only a single case of hepatic decompensation in terms of ascites requiring large-volume paracentesis was observed in one patient during the entire FU. MELD score improved more frequently than Child Pugh score which may be best explained by slight improvements in laboratory values (i.e., bilirubin, creatinine, INR, and albumin), which was sufficient to improve MELD but not Child Pugh score, which categorizes liver function more broadly.

Albeit comprising a shorter FU, increased albumin, stable bilirubin, and no decompensating event or HCC were reported in the study by Degasperi et al., emphasizing the reliability of our results [30]. The BEA score was developed to predict the probability of liver-related events in patients with CHD, and to stratify patients into three different risk categories [17]. In the development cohort, patients with BEA scores B and C had a hazard ratio of 9 and 25, respectively, for liver-related events compared to patients with BEA score A [17]. According to the BEA score, the majority of our patients were categorized as having a moderate and high risk, for whom a higher incidence of hepatic events was anticipated than was actually observed. Although we did not perform survival analysis because of the low patient numbers and relatively short FU for this purpose, our results may suggest that treatment with BLV stabilizes liver function and reduces the risk of liver-related events.

This study has several limitations, which are mainly attributed to the small cohort size and monocentric and retrospective design. On the other hand, our study comprised a relatively long follow-up period (up to 33 months) and a relatively large cohort of patients with CHD and ACLD in comparison to other published real-world experiences with BLV.

In conclusion, treatment with bulevirtide beyond one year appears to be effective and safe in patients with chronic hepatitis D and advanced chronic liver disease. While only one case of hepatic decompensation occurred during a median FU of 26 months, liver function was stable or even improved during treatment in the majority of patients.

Additional Points

Lay Summary. Bulevirtide, which is approved for the treatment of chronic hepatitis D, led to an effective and sustained reduction in hepatitis D viremia in patients with chronic hepatitis D and advanced chronic liver disease. In this retrospective analysis, liver function was stable or improved in most patients without severe adverse effects during long-term treatment with bulevirtide.

Conflicts of Interest

Christoph Schramm received fee for advisory service for Gilead Science GmbH, Germany. All other authors have no conflicts of interest to disclose.