Long-Term Renal Safety Between Patients With Chronic Hepatitis B Receiving Tenofovir vs. Entecavir Therapy: A Multicenter Study

Renal safety is a critical issue in chronic hepatitis B (CHB) patients receiving long-term entecavir (ETV) or tenofovir disofuroxil fumarate (TDF) therapy. We investigated their effects on estimated glomerular ltration rate (eGFR). Treatment-naïve CHB patients receiving ETV or TDF for ≥ 1 year were recruited. The eGFR was assessed using the Chronic Kidney Disease Epidemiology Collaboration equation. We calculated average annual percent change (AAPC) in eGFR using Joinpoint regression.


Introduction
Chronic hepatitis B virus (HBV) infection is a major public health problem affecting approximately 257 million people worldwide [1]. Patients with high serum HBV-DNA levels are more likely to develop liver cirrhosis, hepatic decompensation, and hepatocellular carcinoma (HCC) [2,3]. Thus, suppression of HB-DNA through antiviral therapy (AVT) using oral nucleos[t]ide analogues (NUCs) has become the major basis for treatment of HBV infection [4][5][6][7][8]. Currently, two prominent oral NUCs with high antiviral e cacy and minimal resistance, entecavir (ETV) and tenofovir disoproxil fumarate (TDF), are recommended in the initial treatment of CHB according to the clinical guidelines [4][5][6]. Since oral NUCs can effectively suppress HBV replication, but not eradicate intra-hepatic virus itself, their use should be life-long in most patients. Therefore, along with the issue of antiviral e cacy, the long-term safety of oral NUCs has become a concern.
Of their potential adverse effects, renal safety is the most important given that the medications are known to be excreted by renal ltration. Furthermore, chronic kidney disease is more often associated with patients with CHB than with non-CHB controls [9]. For instance, CHB patients with cirrhosis are more likely to develop renal insu ciency, including acute kidney injury or hepatorenal syndrome, due to reduced renal blood ow and hemodynamic disturbances [10]. Furthermore, renal tubular dysfunction and Fanconi syndrome have been associated with TDF therapy for a decade [11][12][13]. Subsequently, several studies have reported the comparative analyses of renal function between patients receiving ETV and TDF; however, these studies had small sample sizes [14,15] and short follow-up periods [16,17].
Moreover, in most previous studies, renal dysfunction was de ned at a cross-sectional level, that is, only a decrease in the estimated glomerular ltration rate (eGFR) at one or two time points [13,18].
In this multicenter study, we aimed to investigate the long-term effects of ETV and TDF on renal function in a large-scale study population by assessing the serial changes in renal function in the long-term.

Patients
Treatment-naïve CHB patients who received AVT with either daily 0.5 mg ETV (ETV group) or 300 mg TDF (TDF group) as the rst-line regimen between 2007 and 2018 in the Division of Gastroenterology and Hepatology at three academic teaching hospitals in the Republic of Korea (Yonsei University Severance Hospital, Kyungpook National University Hospital, and Cha Bundang Medical Center), were screened for eligibility. The exclusion criteria were as follows: (1) age <19 years, (2) history of HCC at enrollment, (3) decompensated cirrhosis at enrollment, (4) AVT for less than 12 months, (5) co-infection with human immunode ciency virus (HIV) or other hepatitis viruses, (6) history of organ transplantation, and (7) HCC occurrence within 6 months of enrollment (Supplementary Figure 1).
The reimbursement criteria for ETV and TDF in the Republic of Korea are identical (Supplementary Table   1). If histologic information was not available, cirrhosis was clinically de ned according to the following criteria: (1) platelet count <150,000/μL and ultrasonographic ndings suggestive of compensated cirrhosis, including a blunted, nodular liver surface accompanied by splenomegaly (>12 cm); or (2) clinical signs of portal hypertension, such as gastroesophageal varices.
The study protocol complied with the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the institutional review board of each institute. Informed consent was waived because of the retrospective nature of the study.

Clinical evaluation and follow-up
During follow-up, patients underwent routine laboratory tests, including serum creatinine and HBV-DNA levels as well as other viral markers that were monitored at 3-6-month intervals. Patients were screened for HCC and cirrhotic complications every 6 months using ultrasonography and serum alpha-fetoprotein levels [19][20][21][22].
Each patient's renal function was monitored every 3-6 months. The eGFR was calculated using the Data are presented as mean ± standard deviation, or number (%) as appropriate. Differences among continuous and categorical variables were compared using Student's t-test and chi-squared test, respectively. To evaluate the change in eGFR during the 6 years of observation, we used the Joinpoint regression method to calculate the average annual percent change (AAPC) in eGFR, odds ratio (ORs), and 95% con dence interval (CI). Logistic regression analysis was performed to investigate the predictors of ongoing renal dysfunction; ORs and 95% CIs were also calculated. Subsequently, a multivariate regression analysis was performed to assess the independent association between the univariate predictors and ongoing renal dysfunction.

Baseline characteristics
A total of 3,033 patients with CHB who were treated with daily ETV (0.5 mg) or TDF (300 mg) as the rstline antiviral regimen for at least 1 year were analyzed. Table 1 describes the clinical characteristics of these patients at the rst year of AVT. The mean age of the cohort was 48.8 years, and men were predominant (63.8%). Liver cirrhosis was noted in 852 (28.1%) patients, with mean APRI and FIB-4 scores of 0.63, and 2.38, respectively. The mean creatinine and eGFR calculated using the CKD-EPI equation were 0.9 mg/dL, and 92.8 mL/min/1.73 m 2 , respectively. Changes in renal function during AVT among the entire cohort For the entire cohort (n=3,033), eGFR was measured annually for 6 years after the commencement of AVT ( Figure 1).

Changes in renal function according to antiviral regimen
We computed the eGFR at each time point according to treatment with either ETV or TDF ( Figure 2 and Table 2). During the 6-year observation period after the commencement of AVT, the mean eGFR was maintained in the ETV group at 96.0 mL/min/1.73 m 2 in the rst year and 95.6 mL/min/1.73 m 2 in the sixth year (P>0.05). In contrast, the mean eGFR signi cantly reduced in the TDF group from 101.9 mL/min/1.73 m 2 after the rst of AVT to 96.9 mL/min/1.73 m 2 after 6 years (P<0.001). In the ETV group, the mean eGFR was similarly maintained with an AAPC in eGFR of − 0.09% (95% CI, -0.25 to -0.08; P=0.322). In contrast, among the TDF group, the mean eGFR decreased signi cantly every year with an AAPC of -0.88% (95% CI, -1.05 to -0.71; P<0.001)

Predictors of ongoing renal dysfunction
During the 6-year observation period after the commencement of AVT, a total of 681 (22.4%) patients developed ongoing renal dysfunction, which was de ned as a negative slope of the mean change in eGFR during AVT. Table 3 shows the comparison of the characteristics of patients with and without ongoing renal dysfunction. The group with ongoing renal dysfunction had older patients, a higher proportion of those with diabetes mellitus, hypertension, and liver cirrhosis, who received treatment with TDF (all p<0.05), and had lower serum albumin and eGFR at 1 year of AVT, compared to those without.  Abbreviations: , OR, odds ratio; CI, con dence interval; eGFR, estimated glomerular ltration rate; AVT, antiviral therapy; TDF, tenofovir disoproxil fumarate; ETV, entecavir

Discussion
The most recent guidelines from the European Association for the Study of the Liver (EASL) recommend ETV or tenofovir alafenamide (TAF) over ETV in CHB patients with renal insu ciency or its risk factors [4]. Although there have been several studies reporting no difference in renal function between patients treated with TDF and those treated with ETV during a relatively short observation period [14,15,24,25], TDF has the inherent potential to be nephrotoxic. Because TDF accumulates in the proximal tubule during its excretion through the kidney, it causes nephrotoxicity through direct proximal tubular cell and mitochondrial damage [26,27]. In this study, we con rmed that eGFR signi cantly decreased in the TDF group (AAPC -0.88%, P<0.001) during the 6-year observation period after the commencement of AVT, in contrast to the ETV group wherein the eGFR was maintained (AAPC -0.09%, P=0.322).
Our ndings have several strengths. First, we provide robust evidence to support the above EASL guidelines based on real-world data, emphasizing that special attention needs to be paid to serial changes in renal function, especially in the group receiving TDF, given that AVT treatment is usually lifelong [16,17,28,29]. Second, compared to previous studies, our study has the advantage of a longer follow-up duration (up to 6 years) and a larger sample size (n=3,033) from three independent academic teaching hospitals in a real-world setting. Furthermore, from a statistical perspective, we focused on annual changes in the eGFR. In fact, in most previous studies, the occurrence of CKD of or eGFR reduction by ≥20% was adopted as the major study endpoint [16,24,30]. However, given that the annual decline in eGFR among the TDF group, although progressive, was not so rapid, the actual number of events ful lling the occurrence of overt CKD, i.e. eGFR<60 mL/min/1.73 m 2 or eGFR reduction by ≥20% was very small (3.7-6.8% in other studies; 3.7% in our study) [16,24,30]. In such circumstances, there may be a statistically signi cant concern leading to the false negativity of the type II error. For this reason, we focused on the AAPC in eGFR during the long-term AVT, using Joinpoint regression and adopted a negative slope of the mean change in eGFR during the AVT (n=681, 22.4%) as the major clinical endpoint in the logistic regression. To the best of our knowledge, this study is the rst to assess AAPC in eGFR during long-term AVT in patients with CHB.
Notably, in contrast to the ETV group, even patients without diabetes mellitus or hypertension in the TDF group showed signi cantly decreased eGFR, with AAPCs of -0.80% and -0.87%, respectively. As expected, when diabetes mellitus or hypertension was present in the TDF group, the decline in eGFR showed a trend toward more accelerated patterns: an AAPC of -1.59% and -1.00%, respectively. Physicians should monitor patients' renal function more closely, considering that therapy with TDF is almost life-long.
Although patients in the ETV group with diabetes mellitus or hypertension may also experience a signi cant decline in eGFR, this may be due to the natural history of the accompanying comorbidities rather than an ETV-related adverse event. Nevertheless, given that patients with CHB are aging, and their comorbidities (e.g., diabetes, hypertension, and chronic kidney disease) are also increasing [9,31], the early recognition of individual risk factors and correspondingly careful monitoring are also required The TDF-induced eGFR decrease observed in this study occurred consistently every year after the rst year of AVT. There are con icting opinions regarding whether the renal impairment induced by TDF is prominent early since the beginning of AVT [32] or progressive after several years of AVT [13]. Therefore, further research is needed to investigate the timing of the development of TDF-induced renal toxicity, its degree, and annual progress rate. Speci cally, a long-term study of up to 10 years or longer is needed to determine whether the eGFR will continue to drop or there will be a quiescent period. This would provide an important clinical insight into whether TDF should be sustained or replaced with safer oral NUCs that is, ETV or TAF at a certain time point.
An eGFR <60 mL/min/1.73 m 2 and a low albumin level at 1 year of AVT as well as hypertension were independent predictors of ongoing renal dysfunction in our study. Diabetes mellitus was a signi cant univariate predictor, but did not prove to be independently associated with ongoing renal dysfunction in the multivariate analysis; this is different from the general hypothesis that diabetes mellitus is a wellknown risk factor for progression to CKD. This is most likely due to the multi-collinearity of diabetes with hypertension. Indeed, patients with diabetes mellitus were more likely to have hypertension than those without (31.6% vs. 6.9%; OR, 6.20; 95% CI, 4.57-8.42; P<0.001). Consistent with the ndings of previous studies, baseline eGFR <60 mL/min/1.73 m 2 was a poor prognostic factor in our study [24,32], indicating that more caution is required in such a population during follow-up. In addition, albumin synthesis decreases as a result of liver dysfunction in patients with chronic liver disease, and may lead to decreased effective intravascular blood volume, compensatory activation of the renin-angiotensin system and sympathetic nervous system, sodium and water retention, and nally, reduction in renal perfusion and eGFR [33,34].
This study has several limitations. First, the retrospective design may have led to an inherent selection bias, which we tried to overcome by recruiting a homogeneous study population with a statistically reliable large sample size and event number, providing a longer follow-up duration, gathering data from multiple centers (three independent teaching hospital)s, and applying optimized statistical approaches. However, additional studies are required to validate this hypothesis. Finally, the analyses of new biomarkers for CHB (e.g., serum quantitative hepatitis B surface antigen, serum hepatitis B core-related antigen, serum HBV-RNA, or speci c HBV mutants), and for renal insu ciency (i.e., cystatin C, β2microglobulin, or neutrophil gelatinase-associated lipocalin) may provide more precise predictions in the future [13,35].
In conclusion, compared to the ETV group, the TDF group experienced slow, but progressive renal dysfunction. Although the annual eGFR change was small in the TDF group, careful monitoring of renal function is necessary to allow timely intervention, especially in patients requiring life-long AVT.

Declarations
Data availability: Available on request to corresponding author.