Real‐world safety and effectiveness of tenofovir alafenamide for 144 weeks in Japanese patients with chronic hepatitis B

Tenofovir alafenamide (TAF), a prodrug of tenofovir, delivers high levels of active drug to hepatocytes and is given in a lower dose than tenofovir disoproxil fumarate (TDF). TAF reduces viral replication in patients with chronic hepatitis B (CHB) similar to TDF and has shown a lower risk of the renal and bone toxicities associated with TDF use. This post‐marketing surveillance study examined the safety and effectiveness of TAF in treatment‐naïve and ‐experienced CHB patients who received TAF for 144 weeks at real‐world clinical sites in Japan. Safety assessments included the incidence of adverse drug reactions (ADRs), renal and bone events, and changes in selected laboratory parameters. Effectiveness was based on the proportion of patients with HBV DNA levels below the lower limit of quantitation or <29 IU/mL. This analysis included 580 patients; 18.4% of whom were treatment‐naïve. The cumulative incidence of ADRs was 0.21 per 100 person‐months, and the incidence of serious ADRs was 0.01 (95% CI, 0.00–0.04) per 100 person‐months. There were no ADRs of declines in estimated glomerular filtration rates, renal failure or proximal tubulopathy. The most common ADR was hypophosphataemia in seven (1.2%) patients. Two (0.4%) patients each had decreased blood phosphorus, bone mineral density decreased, dizziness and alopecia. Overall, the proportion of virologically suppressed patients increased from 68.8% at baseline to 97.5% at Week 144. These results confirm the real‐world safety and effectiveness of TAF in Japanese patients with CHB and are consistent with the findings of other evaluations of the safety and efficacy of TAF in CHB.

year. 1 Recognizing the profound public health implications of CHB, the WHO has established a global health strategy to eliminate HBV infections by 2030.A principal component of this strategy is ensuring widespread access to the most effective therapies for CHB. 2 Nucleos(t)ide analogues (NAs) are the mainstay of antiviral treatment for CHB due to their ability to achieve a sustained suppression of HBV replication, which is essential for the normalization of serum alanine transaminase (ALT), HBeAg loss and improvements in liver histology. 3Randomized, controlled trials have established entecavir (ETV), [4][5][6][7] tenofovir disoproxil fumarate (TDF) 8 and tenofovir alafenamide (TAF) [9][10][11][12] as highly effective therapies for HBeAg-positive and HBeAg-negative patients.However, antiviral drug resistance is an important determinant of the success of longterm CHB therapy. 13,14[30][31] Furthermore, recent The Japan Society of Hepatology (JSH) Guidelines recommend switching patients from TDF to TAF, including good responders, if there are concerns about long-term safety, and switching from ETV to TAF in patients at risk of resistance. 30F is a novel phosphonamidate prodrug of tenofovir that offers several advantages over TDF, including greater plasma stability [32][33][34] and high levels of active drug (TFV diphosphate) delivered to HBVinfected hepatocytes at a lower dose than TDF. 357][38][39][40] Long-term surveillance in real-world patient populations is important to determine whether the favourable effects seen in controlled trials of TAF are maintained in clinical practice, including demonstrations of no new safety signals.
The current post-marketing surveillance study, which was undertaken as a requirement of the Pharmaceuticals and Medical Devices Agency of Japan, determined the incidence of adverse drug reactions (ADRs) during 144 weeks of TAF in CHB patients who received care in real-world clinical sites in Japan.A secondary objective was to evaluate the effectiveness of TAF in treatment-naïve and -experienced patients who were followed at these sites.

| Study population
Patients were eligible for this study if they initiated TAF for the treatment of CHB according to the Japan package insert 41 at local clinical practices in Japan.This included both treatment-naïve patients and those who had previously received other medications approved for CHB (treatment-experienced), such as TDF, ETV, adefovir dipivoxil (ADV) and lamivudine (LAM).The study received approval by the institutional review boards at each participating site, complied with all relevant regulations of the Pharmaceuticals and Medical Devices Agency and followed the principles of the International Conference on Harmonization Pharmacovigilance Planning E2E guidelines, Good Post-marketing Study Practice and Good Pharmacovigilance Practices.

| Study design
This multicentre, observational study prospectively enrolled patients from 91 clinical sites throughout Japan between July 2017 and August 2021.Each site used a commercial, web-based, electronic data capture system (PostMaNet™; Fujitsu FIP Corporation) for data collection and management.Data were collected on electronic case report forms (eCRFs) at baseline and pre-specified follow-up intervals at 24, 48, 96 and 144 weeks.Variables of interest included demographics, clinical characteristics and treatment status at baseline; laboratory evaluations (haematology and clinical chemistries); HBV assessments (e.g.serologies and HBV DNA levels); resistance development; adverse events; and other observations such as BMD and liver fibrosis status at baseline and during treatment.The observation period for safety and efficacy was enrolment through 144 weeks or 4 weeks after treatment discontinuation (and the start of other HBV therapies), whichever came first.

| Objectives
Evaluation of the primary safety objective was based on the frequency of ADRs during the study period.The time to first onset of ADRs was categorized as <24 weeks, 24 to <48 weeks, 48 to <72 weeks, 72 to <96 weeks, 96 to <120 weeks, 120 to <144 weeks and ≥ 144 weeks.
Additional safety outcomes included the occurrence of post-treatment hepatic flares; renal toxicity; bone events/loss of BMD; severe hepatomegaly with lactic acidosis and steatosis; pancreatitis, redistribution or accumulation of body fat; and long-term safety information on adults, which were evaluated as pre-specified safety risks.Changes in BMD were assessed as per the standard of care at each site and could include dual-energy X-ray absorptiometry of the hip and lumbar spine, with test results compared to the young adult mean (YAM) BMD.
Laboratory evaluations, performed at local facilities, included ALT, aspartate aminotransferase (AST) and estimated glomerular filtration rate (eGFR).The calculated values of eGFR were based on serum creatinine, age and gender using the Chronic Kidney Disease Epidemiology Collaboration method (eGFR CKD-EPI ). 42Stage 1 chronic kidney disease (CKD) was defined as eGFR CKD-EPI ≥ 90 mL/min/1.73m(severe) as 15-29 mL/min/1.73m 2 and Stage 5 (end stage) was defined as <15 mL/min/1.73m 2 .The primary measure of effectiveness was the proportion of patients with HBV DNA levels reported by local PCR assays that were less than the lower limit of quantification (LLOQ) or less than 29 IU/ mL at 144 weeks after the start of TAF.Additional effectiveness measures were the proportion of patients with ALT normalization in those with baseline ALT levels that exceeded the upper limit of normal (ULN) and the proportion of patients who experienced HBeAg (HBeAg-positive patients only) and HBsAg loss and seroconversion (all patients).The ULN for ALT was defined as 30 IU/L according to the JSH Guidelines for the Management of Hepatitis B Virus Infection (version 2.2; May 2016).

| Statistical analysis
An ADR was defined as an untoward medical occurrence considered causally related to TAF at any dose administered.In the event of two determinations for causality from the physician report and the sponsor report, the more conservative assessment was accepted.
ADRs were coded by preferred terms using MedDRA/J version 24.0.
Events with the same preferred term were counted once if they occurred multiple times in the same patient.The number and incidence of ADRs were calculated as the number of patients with an ADR divided by the total number of patients in the safety analysis set.The overall ADR incidence rate and two-sided 95% confidence intervals (CIs) were calculated as the number of new events per 100 personmonths at risk.The cumulative frequencies of ADRs were compared between subgroups by patient characteristics using Chi-square tests with a significance level of 5% or a Fisher's exact test if the expected frequency for a specific ADR was <5.
The proportion of patients with HBV DNA results that were below the LLOQ and two-sided 95% CI was calculated for each follow-up assessment.The denominator for the baseline proportion was the number of patients in the effectiveness analysis set, including those with unknown or not described results.The denominator for the proportion at each of the follow-up assessments was the number of patients in the effectiveness analysis set excluding those with results reported as unknown or not described.The proportions of patients with HBV DNA negative results were compared between subgroups by patient characteristics using Chi-square tests with a significance level of 5% or Fisher's exact tests in instances where the expected frequency was <5.

| Patient disposition
A total of 581 patients from 91 sites in Japan were enrolled, and eCRFs were completed for 580 patients (Figure 1).The safety analysis set included 577 of the 580 patients with completed eCRFs and at least one administered dose of TAF; three patients were excluded because there were no recorded visits after the date of the first TAF prescription.The effectiveness analysis set consisted of 576 patients from the safety analysis set, with one patient excluded due to no reports on effectiveness outcomes following treatment initiation.

| Patient characteristics at baseline
The mean (standard deviation [SD]) age of patients in the safety analysis set was 58 ( 13) years (Table 1), and 34.7% were 65 years or older.The majority of patients were male (60.7%), and TAF was initiated in an outpatient setting for 97.8% of patients.Prior treatment was reported for 81.6%, and 18.4% of patients were treatmentnaïve.Of the 471 treatment-experienced patients, 98.5% had been treated with NAs; TDF was used in 56.9% and ETV in 54.5%.
A past medical history for any hepatic condition, defined as conditions that were previously diagnosed and noted by the investigators not to be ongoing at the start of TAF treatment, was reported for 10.6% of 577 patients in the safety analysis set (Table 2).HCC was the most common past medical condition, which was diagnosed in 8.7% of patients.A current hepatic comorbidity was reported for 126 (21.8%) patients, with cirrhosis the most frequently reported condition in 14.4%.Among patients with cirrhosis, 89.2% had compensated cirrhosis, 9.6% had decompensated cirrhosis, and status was unknown or not reported for one patient (1.2%).Other current hepatic comorbidities included HCC in 3.8% of patients, chronic liver disease of non-HBV aetiology in 2.1% and some other type of hepatic comorbidity in 6.1% of patients.
Concurrent non-hepatic comorbidities were reported for 51.1% of the safety analysis set.The most frequently reported comorbidities were treated hypertension in 22.2% of patients, hyperlipidaemia in 12.8%, diabetes in 9.9%, renal impairment in 9.0% and other cardiovascular conditions in 6.1% of patients.Other current comorbidities were reported for approximately one-third of patients.
At baseline, 59.3% of 577 patients had Stage 1 CKD, 25.5% were Stage 2, 5.9% Stage 3A or 3B and 0.2% had Stage 4 CKD.HBV DNA levels were < 3 log 10 IU/mL in 437 (75.7%) patients.Among patients with HBV DNA >3 log 10 IU/mL, levels were 3 to <7 log 10 IU/mL and ≥ 7 in 10.7% and 4.3% of patients, respectively.HBV DNA level was unknown or not reported for 9.2% of patients.Baseline HBe antigen (HBeAg) results were negative in 42.1% of patients and positive in 19.2% of patients, and HBe antigen status was unknown or not reported in 38.7% of patients.ALT levels at baseline were ≤ULN in 63.1% of patients, >ULN in 29.5%, and unknown or not reported for 7.5% of patients.

| Treatment characteristics
TAF was discontinued in 11.6% of 571 patients (Table 3).The reported reasons for discontinuation were loss to follow-up (n = 39), adverse events (n = 16) and switch to another treatment regimen (n = 8).Only one patient discontinued TAF due to lack of effectiveness, and genotypic drug resistance sequence testing was not performed for this individual.Treatment discontinuation due to death was reported for six patients, including HCC in three patients and one patient each with hepatic cancer, chronic hepatic failure and metastases to lung and sudden death.None of the fatal events was reported as an ADR.

| Safety
In the safety analysis set, 37 ADRs were reported in 29 patients for an overall incidence rate of 0.21% (95% CI, 0.15-0.28)per 100 person-months.The most common ADR was hypophosphataemia in seven (1.2%) patients (Table 4).Additionally, two (0.4%) patients each had ADRs of decreased blood phosphorus, decreased bone density, dizziness and alopecia.Twenty-seven of 31 (87.1%)first-onset ADRs were reported by 96 weeks following initiation of TAF, and the incidence rates for ADRs were comparable over time.
Two serious ADRs (SADRs) were reported in two patients for an overall incidence rate of 0.01 (95% CI, 0.00-0.04)per 100 personmonths.The SADRs were hepatic function abnormal and idiopathic interstitial pneumonia in one patient each.The patient with serious hepatic function abnormality had a concurrent comorbidity of hepatic steatosis and ALT levels of 200 IU/L at baseline and 347 IU/L at the onset of the event (1 month after TAF initiation); bilirubin values were normal.
Administration of TAF was continued, ALT values were 46 IU/L and 30 IU/L after 2.5 and5 months after the event onset, respectively, and the outcome was reported as resolved.The patient who experienced idiopathic interstitial pneumonia had a history of idiopathic interstitial pneumonia.Additional details for this event were not reported, including the outcome of the event and if TAF was discontinued.

| Bone safety
Two (0.4%) patients experienced non-serious ADRs of decreased bone density.These patients continued TAF, with outcomes reported as recovering for one patient and unknown in the second patient.
a Regarding the single patient who was initially enrolled without any CRF collection, the site later identified in their internal check that written informed consent from the patient could not be confirmed.Therefore, adhering to the site's internal policy, a request was issued to not proceed with CRF data collection.

| Other safety specifications
There were no ADRs related to hepatic flares during treatment; lactic acidosis and severe hepatomegaly due to hepatic steatosis; pancreatitis; or redistribution or accumulation of body fat.When there were multiple reasons for discontinuation, each reason was counted separately and patients were counted more than once.The number of patients with a 'yes' response for treatment discontinuation was used as the denominator.

TA B L E 4
Frequency and type of ADRs: Safety analysis set (N = 577).

| DISCUSS ION
In this multicentre, real-world observational study that included 580 patients who were prospectively enrolled from 91 sites across Japan between July 2017 and August 2021, the favourable safety and effectiveness profile of TAF that has been demonstrated in clinical trials 10,11 and real-world analyses [36][37][38][43][44][45]  Taken together, these findings confirm the excellent safety profile of TAF when administered to community-dwelling adults with CHB regardless of prior treatment status and regimen, comorbidities, and older age.
The risk of renal toxicity is an important clinical concern for CHB patients, especially in older patients and those with comorbid health conditions that affect kidney function-conditions that were well represented in our study population.Renal toxicity has been an issue of particular concern for CHB patients who require long-term therapy with TDF or ADV.A retrospective, cohort study found that during 3 years of TDF treatment, the proportion of patients (n = 426) who experienced renal impairment was 2.9% at Year 1, 1.8% at Year 2 and 1.7% at Year 3. 46 Similarly, a retrospective, non-interventional study compared the real-world renal safety and effectiveness of TDF versus ETV.Among patients treated with TDF, 1.9% had renal tubular dysfunction and 7.1% had renal adverse events that resulted in discontinuation of TDF while no patients treated with ETV experienced these events. 12ere is growing evidence that TDF-induced renal toxicity can be reversed without a loss of efficacy when patients are switched to TAF.A Phase 3, double-blind, non-inferiority study of patients who were randomly assigned to switch to TAF or remain on TDF found a significantly smaller change in creatinine clearance and improved tubular function in patients treated with TAF compared to those who continued TDF.Importantly, TAF was non-inferior to TDF with respect to clinical efficacy. 45These findings were confirmed in a narrative review of 36 papers that examined efficacy and safety data in patients who were switched from TDF to TAF.Notably, in eight studies that reported renal safety outcomes, switching from TDF to TAF was associated with improved renal parameters.Efficacy was similar between TAF and TDF. 47 the current study, two patients experienced non-serious ADRs of decreased bone density, although only a small number of patients underwent BMD testing, which limits any definitive conclusions about these results.The low occurrence of changes in BMD in this study, however, is consistent with results from Phase 3 global studies that reported significantly smaller declines in hip and spine BMD at 48 weeks in HBeAg-positive and HBeAgnegative patients who switched from TDF to TAF. 10,11 Collectively, these results taken with those on renal ADRs confirm a favourable renal and bone safety profile for TAF that might be attributed to lower plasma concentrations and reduced exposure for the kidneys and bones. 48,49ere were no reports of AEs attributed to changes in lipid levels in the current study, which is consistent with findings of other real-world studies. 37,50However, lipid changes may not have been routinely monitored by hepatologists who participated in this study.
While a clear association of switching from TDF to TAF on the development of atherosclerosis has not been well established, 51 it has been suggested that clinicians consider regular monitoring of fasting lipids in CHB patients taking TAF, particularly those who are at risk of age-related comorbidities such as atherosclerosis. 44ter 144 weeks of TAF, high rates of HBV DNA negativity were achieved (approximately 98% overall), with no differences in viral suppression rates seen with clinical characteristics such as renal impairment, cirrhosis or prior treatment with NAs.These findings are consistent with those observed in clinical trials 10,11 as well as real-world settings 36,38,43 and offer further support for TAF as an alternative to TDF in the long-term management of CHB patients, especially those with renal impairment and those at higher risk for renal or bone diseases.
Our results are of particular relevance to CHB patients in Japan, where the estimated HBsAg seroprevalence is 1.02% (95% CI, 1.01-1.02), 52and a disproportionate number of affected individuals are of advanced age and may have comorbid health conditions such as renal impairment.It has recently been estimated that only 32% of treatment-eligible patients in Japan actually receive antiviral therapy, 53 which suggests that many patients may be unaware that they are infected or require treatment.The Basic Act on Hepatitis Measures establishes as a priority that all persons in Japan have equal access to testing and appropriate treatment for CHB. 54lative to other NAs, TAF offers patients a more favourable bone and renal safety profile with comparable long-term effectiveness and may have an important role in ensuring that Japanese patients receive optimal treatment.
The current study has several limitations.The analysis presented in this manuscript follows what was outlined in the study's statistical analysis plan and presented to the Pharmaceuticals and Medical Devices Agency (PMDA).Most notable of these is the single-arm study design with no comparator arm, which is a feature of many real-world analyses.The study relied on data obtained from clinical settings that reflected the treatment patterns for CHB at community-based practices located throughout Japan, which limits the generalizability of these results.However, patients were representative of those seen in usual community settings and were diverse with respect to their medical history, the presence of comorbidities and experience with CHB therapies.Further, safety and efficacy assessments were rigorous and consistent across the participating practice sites.
The reason patients were switched to TAF following treatment with other approved therapies for CHB was not collected.This information would help identify factors that prompted the decision to switch treatment, which might include channelling bias or practice patterns and preferences.This study also did not evaluate the safety and effectiveness of TAF in special populations such as pregnant women or immunosuppressed patients.However, this study assessed the use of TAF according to the label in Japan, and it was expected that these populations would not be included in the current study.Additional research is needed to clarify the role of TAF in the management of these patients.

| CON CLUS IONS
This real-world post-marketing surveillance study conducted in Japan further confirms that TAF offers a safe and effective treatment for CHB, which is consistent with clinical trial and real-life evaluations of TAF in the management of CHB.No new safety or effectiveness issues were identified, and no additional safety or effectiveness measures are considered necessary.

4 |
HBsAg loss and seroconversionOf 195 patients who were HBsAg positive and negative or unknown status for HBsAb negative at baseline and also had results at Week 144, one patient experienced HBsAg loss (0.5% [95% CI, 0.01%-2.8]);no patient had HBsAg seroconversion.
Demographic and baseline characteristics (N = 577).Current and past hepatic comorbidities, current non-hepatic comorbidities and laboratory assessments at baseline (N = 577).Frequency of and reasons for treatment discontinuation.