Potential conflicts of interest: Dr. Nguyen H. advises Bristol-Myers Squibb. He also advises and is on the speakers' bureau of Gilead. Dr. Garcia is a consultant for, advises, and received grants from Bristol-Myers Squibb. He is a consultant for, advises, is on the speakers' bureau of, and received grants from Gilead. He also received grants from Novartis and Roche. Dr. Nguyen is a consultant for, advises, and received grants from Bristol-Myers Squibb and Gilead. She also received grants from Roche and Novartis. Dr. Trinh owns stock in, is a consultant for, advises, is on the speakers' bureau of, and received grants from Gilead. He is a consultant for, advises, is on the speakers' bureau of, and received grants from Bristol-Myers Squibb. He also received grants from Roche.
Renal dysfunction has been reported in patients treated with adefovir dipivoxil (ADV); however, its incidence and clinical importance may be underappreciated given the lack of long-term follow-up and data outside of a clinical trial setting. Our goal was to examine the severity and incidence of renal dysfunction in a real-life setting for patients treated with ADV and whose baseline estimated glomerular filtration rate (eGFR) was >50 mL/minute. We performed a cohort study of 290 chronic hepatitis B patients: 145 patients treated with 10 mg ADV and 145 patients unexposed to ADV at two community clinics, who were matched for age (±10 years), sex, and baseline eGFR. The exposed and unexposed populations were well-matched with a similar mean age (46–47 years), proportion of male patients (76.5%), baseline serum creatinine (0.97–0.99 mg/dL), and baseline creatinine clearance (85.0–85.4 mL/minute). The incidence density for renal dysfunction defined by treatment termination and/or development of eGFR ≤50 mL/minute was five cases per 100 patient-years in the exposed group compared with 1.36 cases per 100 patient-years in the unexposed group (P = 0.02). The relative risk of exposed to unexposed was 3.68 (95% confidence interval 1.1–19.3). On Cox proportional hazard analysis also inclusive of sex, ADV was a significant predictor of significant renal dysfunction (hazard ratio [HR] 3.94, P = 0.03). There were also significant trends for age >50 years (HR 3.49, P = 0.087), mild renal impairment at baseline (HR 4.49, P = 0.073), and hypertension and/or diabetes mellitus (HR 2.36, P = 0.074). Conclusion: ADV is an independent predictor for significant deterioration of renal function. Patients on ADV should be monitored, especially patients who are older, have baseline renal insufficiency, or have hypertension and/or diabetes mellitus. (HEPATOLOGY 2009.)
Chronic hepatitis B (CHB) affects more than 350 million people worldwide and remains a major public health problem.1, 2 Approximately 15%–40% of infected individuals will develop chronic liver disease leading to cirrhosis and hepatocellular carcinoma, which account for half a million deaths annually worldwide.2, 3 Current practice guidelines include treatment with antiviral therapy for patients who may be at higher risk for end-stage liver disease and hepatocellular carcinoma.4 The goal of therapy is to eliminate or significantly suppress hepatitis B virus (HBV) replication to prevent disease progression, major complications, and mortality.4, 5
There are currently seven therapeutic agents approved in the United States for the treatment of CHB. Injection therapy includes standard and pegylated interferon-α. Oral nucleoside analogues include lamivudine (LAM), entecavir (ETV), and telbivudine. Oral nucleotide analogues include adefovir dipivoxil (ADV) and tenofovir. The high incidences of severe side effects and/or low rate of response of standard and pegylated interferon-α have limited their use. An unacceptably high incidence of drug resistance with long-term lamivudine therapy has made it a poor option for most patients given the availability of newer and improved agents.6–8
ADV is a nucleotide analogue of adenosine monophosphate that is effective in viral suppression in both treatment-naïve and lamivudine-resistant CHB patients.9, 10 ADV is eliminated renally through glomerular filtration and active tubular secretion.11, 12 Long-term use of ADV has been reported to be associated with a small risk of nephrotoxicity.13 Nephrotoxicity has been listed as a warning associated with the medication, and it is recommended that renal function of all patients treated with ADV be monitored closely.14 Incidences of nephrotoxicity at 48 weeks were reported in 13%, 27%, and 50% for ADV doses of 30 mg, 60 mg, and 120 mg, respectively.11 Renal impairment has also been reported for doses of 10 mg and was generally defined as an increase in serum creatinine (Cr) level by 0.5 mg/dL, as confirmed by two consecutive laboratory results in clinical trials involving this medication.3 However, serum Cr level alone is not an accurate measurement of renal function. A given serum Cr level may represent different levels of Cr clearance and renal function depending on the patient's weight, age, and race; therefore, such definition of renal impairment as used in clinical trials of ADV may underestimate the incidence of clinically significant renal impairment. In addition, long-term follow-up in registration trials is often limited, and patients enrolled in such trials are also usually carefully selected and may be different from the general CHB patient population encountered by clinicians in a real-life setting.3
According to the National Kidney Foundation practice guideline for chronic kidney disease, serum Cr concentration alone should not be used to assess the level of kidney function.15 Evaluating patients' estimated glomerular filtration rate (eGFR) level using the Cockcroft-Gault formula [(140 − age) × (weight in kilograms) × (0.85 if female) / (72 × Cr)] yields the best indication of current kidney function,15 because it accounts for Cr level, weight, age and sex.
Few studies to date have examined eGFR fluctuation in CHB patients treated with ADV, especially in Asian patients who generally have smaller body mass, and smaller changes in serum Cr may represent a higher degree of renal function change. Safety data available for a 10-mg dose only report fluctuation in serum Cr level ≥0.5 mg/dL. The aim of this study was to evaluate the incidences and magnitude of decrease in eGFR of Asian American patients with CHB treated with 10 mg ADV.
We performed a retrospective cohort study of 145 consecutive CHB patients who were started on 10 mg ADV daily between January 2002 and January 2006. All 145 patients were treated by five gastroenterologists in two community-based clinics in the San Francisco Bay area in California. The exposed group consisted of patients who were prescribed 10 mg ADV and had a baseline eGFR ≥50 mL/minute. Patients were excluded if their baseline eGFR was <50 mL/minute.
A matched 1:1 unexposed population was randomly selected and consisted of 145 CHB patients who were started on ETV from January 2005 and June 2008. The unexposed group consisted of patients who were treated with 0.5 or 1.0 mg ETV and had a baseline eGFR ≥50 mL/minute. Patients were excluded if they had been exposed to ADV prior to treatment with ETV and/or had a baseline eGFR <50 mL/minute. The exposed and unexposed group were matched by their baseline characteristics for age (±10 years), sex, and baseline eGFR classification. We defined exposed as having exposure to ADV and unexposed as having no exposure to ADV but exposure to another anti-HBV therapy.
All study patients were treated with either ADV or ETV. Patients receiving other forms of antiviral therapy for CHB were not included in the study as they make up a very small minority (<5%) of all patients treated for CHB in these clinics. The inclusion of patients treated with one single alternative agent as the group unexposed to ADV was chosen to avoid unnecessary heterogeneity in this comparison group. Patients not treated with any anti-HBV medications were not chosen for the comparison group because they do not receive frequent laboratory testing to allow for close monitoring of renal function.
We queried and compiled a list of 383 patients on ETV and abstracted only their age, sex, and baseline serum Cr level. Matching of the unexposed group was based only on selected factors. Data abstraction was not performed until patients were matched with patients in the 10 mg ADV group. CHB patients receiving another anti-HBV medication, such as ETV, were selected as the comparison group so that they would most likely have similar baseline CHB and frequent laboratory testing as the exposed group to allow for assessment of their renal function. Pegylated interferon-α2a, lamivudine, and telbivudine were rarely used in our patient population due to tolerability and/or efficacy issues. Tenofovir was also used in a small number of our patients because it is only recently approved; its choice as a comparison group for ADV should also be avoided, because it belongs to the same class of nucleotide analogs.
We then followed and compared the incidence density of renal dysfunction in the exposed group versus the unexposed group. Exposed time was measure in person-years. Date of entry was determined by baseline date, and date of exit was defined by treatment termination or termination of study.
The study was approved by the institutional review board at Stanford University (Stanford, CA).
All available medical records were evaluated thoroughly and data were recorded using a case report form developed for this study. Patient eGFRs were calculated using the Cockcroft-Gault formula. Baseline eGFR classifications are based on the manufacturer packet insert and are as follows: unimpaired (eGFR ≥ 80 mL/minute), mildly impaired (50 mL/minute ≤ eGFR < 80 mL/minute), moderately impaired (30 mL/minute ≤ eGFR < 50 mL/minute), and severely impaired (eGFR < 30 mL/minute).14 We also performed a subanalysis of baseline and changes in eGFR using the Modification of Diet in Renal Disease (MDRD) formula [170 × (Cr−0.999) × (age−0.176) × (0.762 if female) × (1.180 if African American) × (blood urea nitrogen−0.170) × (albumin0.318)], which takes into account changes in albumin and ethnicity though both of these values are expected to be rather stable or homogeneous in our population.16 Termination of treatment was determined by patients' treating physicians and patients were generally switched to an alternative anti-HBV medication.
Laboratory tests were performed at one of the local community clinical laboratories operated by either Quest Diagnostics (San Juan Capistrano, CA) or LabCorp (Burlington, NC). Cirrhosis was defined by the presence of stage 4 fibrosis on histology or by the presence of portal hypertension (platelet count <120,000, splenomegaly, ascites, encephalopathy, or varices).
Descriptive statistics were reported as proportion (%) for categorical variables, and mean ± standard deviation or median (range) for continuous variables. Categorical variables were evaluated using the chi-square test. Normally distributed continuous variables were evaluated using the Student t test. Continuous variables that were not normally distributed were evaluated using nonparametric methods. The Cox proportional hazard regression model was used to estimate univariate and multivariate hazard ratios (HR) relating sex, age, presence of hypertension and/or diabetes mellitus, exposure to ADV, and baseline eGFR classifications to development of renal dysfunction (decreases in eGFR and development of moderate renal dysfunction during treatment). The Kaplan-Meier method was used to calculate the proportion of patients who maintain their renal function during treatment for CHB in both treatment cohorts, and comparison between groups was evaluated using the log-rank test. The 2 × 2 method was used to calculate the incidence density of renal dysfunction in the exposed and unexposed groups. The comparison of incidence rate between groups was performed using the Fisher's exact test. Statistical significance was defined with a two-tailed P-value ≤ 0.05. All statistical analyses were performed using Stata 9.1 (Stata Corporation, College Station, TX).
Baseline Clinical Characteristics.
A total of 145 consecutive CHB patients prescribed ADV between January 2002 and January 2006 (exposed cohort) and 145 consecutive CHB patients prescribed ETV between January 2005 and June 2008 (unexposed cohort) were included in the data analysis.
Baseline characteristics for both groups are shown in Table 1. The exposed and unexposed populations were well-matched with a similar mean age, similar proportion of male patients, and similar proportion of baseline eGFR classification. Patients who were hepatitis B e antigen–positive made up 29% of the ADV group and 43.5% of the unexposed group. Baseline HBV DNA level in the ADV group was lower than in the unexposed group (5.0 ± 2.3 versus 6.1 ± 2.0 [P = 0.0001]). Both the ADV and unexposed group had similar baseline total bilirubin levels (0.8 ± 0.5 versus 0.8 ± 0.3 [P = 0.27]) and albumin levels (4.3 ± 0.5 versus 4.4 ± 0.4 [P = 0.42]). The ADV and unexposed groups also had a similar proportion of patients with hypertension (27.6% versus 24.1% [P = 0.50]) and diabetes mellitus (12.4% versus 9% [P = 0.34]).
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; HBeAg, hepatitis B e antigen.
Values are expressed as the mean ± standard deviation, median (range), or number of patients (%).
Age, years (n = 145/145)
46.7 ± 11.8
46.2 ± 11.6
Male sex (n = 145/145)
Ethnicity (n = 145/145)
Weight, kg (n = 145/145)
65.2 ± 10.7
63.8 ± 9.6
Treatment duration, months
HBeAg-positive (n = 145/145)
ALT, U/L (n = 145/145)
AST, U/L (n = 144/143)
Total bilirubin, mg/dL (n = 139/142)
0.8 ± 0.5
0.8 ± 0.3
Albumin, g/dL (n = 144/142)
4.3 ± 0.5
4.4 ± 0.4
HBV DNA, log10 IU/mL (n = 145/145)
5.0 ± 2.3
6.1 ± 2.0
Cirrhosis (n = 145/145)
HTN (n = 145/145)
DM (n = 145/145)
Baseline and Changes in Renal Function During Treatment for CHB.
Baseline renal indices for both groups are listed in Table 2. Baseline serum Cr and eGFR were comparable in both groups. Mean baseline serum Cr levels and eGFR in the ADV and unexposed groups were well-matched. There was a statistically significant difference in the number of incidences and severity of decrease in eGFR in the ADV group versus the unexposed group (Fig. 1). A mild decrease of eGFR (10%–20%) from baseline was observed in a total of 30% in the ADV group versus 16% among the unexposed group during follow-up, a moderate decrease (20%–30%) was observed in 15% versus 6%, and a severe decrease (≥30%) was observed in 7% versus 1% (P = 0.0001). We found similar trends for the number of incidences and severity of decrease in eGFR in the ADV group versus the unexposed group using the alternative calculation for eGFR with the MDRD formula (P = 0.0001). There was also a statistically significant difference in the incidence of elevated severity classification of renal function (Fig. 2). Of the patients on ADV with unimpaired renal function, 18 (22.7%) developed mildly impaired renal function (IRF) and one (1%) developed moderate IRF compared with 12 (15.2%) and 0 (0%) in the unexposed group (P = 0.005). Of the patients on ADV with mild IRF at baseline, 16 (24.2%) developed moderate IRF compared with three (4.5%) in the unexposed group (P = 0.005). A statistically significant difference was also observed in the incidence of elevated severity classification of renal function between the ADV group versus the unexposed group using the MDRD formula (P = 0.0001). Three (2.1%) patients in the unexposed group developed moderate renal dysfunction compared with 17 (11.7%) in the ADV exposed group. A total of seven patients (5%) had an increase in serum Cr of ≥0.5 mg/dL in the exposed group.
Table 2. Renal Function of ADV Patients and Unexposed Patients: Baseline Characteristics and Outcomes
Patients reclassified to a higher category of renal impairment classification
Patients reclassified to moderate renal impairment classification
Patient switched off treatment due to renal impairment
Discontinuation of treatment was observed in 10 (6.9%) patients on 10 mg ADV compared with 0 (0%) patients in the unexposed group (P = 0.004). Detailed patient characteristics for those who were switched off ADV are described in Table 3. All 10 patients recovered from renal impairment with their eGFR returning to baseline (Fig. 3) with a mean recovery period of 9 ± 4 months. One patient was reintroduced to 10 mg ADV after being switched to an alternative anti-HBV therapy, and this patient experienced a relapse of renal dysfunction.
Table 3. Characteristics of Patients Switched Off ADV Due to Renal Dysfunction
Baseline HBV DNA log10IU/mL
Baseline eGFR (mL/minute)
Lowest eGFR During Treatment
Months on Treatment Before Switching
eGFR Returned to Baseline Postswitch
Abbreviations: HBeAg, hepatitis B e antigen; M, male; NA, not applicable.
Figure 4 shows that the proportion of patients who maintained their eGFR within 20% from baseline was also significantly lower in the ADV group when compared with the unexposed group using the Kaplan-Meier method (P = 0.0104).
Incidence Density for Renal Dysfunction.
The incidence density for renal dysfunction defined by treatment termination and/or eGFR ≤50 mL/min was five cases per 100 patient-years in the exposed group compared with 1.36 cases per 100 patient-years in the unexposed group (P = 0.02). The relative risk of exposed to unexposed was 3.68 (95% confidence interval 1.1–19.3).
Predictors for Significant Decrease in eGFR.
Table 4 describes the univariate and multivariate predictors and hazard ratios for incremental decrease in eGFR of ≥20%. On multivariate analysis, following adjustments for presence of hypertension (HTN) and/or diabetes mellitus (DM), independent predictors of ≥20% decrease in eGFR were treatment with 10 mg ADV therapy (HR 2.30, P = 0.018), age >50 (HR 3.02, P < 0.004), presence of mild renal impairment at baseline (HR 0.43, P = 0.021), and male sex (HR 0.46, P = 0.014).
Table 4. Predictors for Decrease in eGFR ≥20% from Baseline
Model inclusive of age, sex, presence of HTN and/or DM, presence of mild renal impairment at baseline (50 mL/minute ≤ eGFR ≤ 80 mL/minute), and therapy with ADV.
Age >50 years
HTN and/or DM
Mild renal impairment at baseline
Table 5 describes the predictors for development of moderate renal impairment and/or treatment discontinuation due to renal impairment and their hazard ratios. A significant independent predictor was 10 mg ADV therapy (HR 3.94, P = 0.03) following adjustment for sex. There were also trends for age >50 (HR 3.49, P = 0.087), presence of HTN and/or DM (HR 2.36, P = 0.074), and presence of mild renal impairment at baseline (HR 4.49, P = 0.073) as predictors for moderate renal impairment and/or treatment discontinuation.
Table 5. Predictors for Moderate Renal Impairment and/or Treatment Termination
Model inclusive of age, sex, presence of HTN and/or DM, presence of mild renal impairment at baseline (50 mL/minute ≤ eGFR ≤ 80 mL/minute), and therapy with ADV.
Age >50 years
HTN and/or DM
Mild renal impairment at baseline
Although there are many alternative treatments for CHB, ADV (10 mg) remains one of the more common therapies, because it was the first form of nucleotide analogue available for treatment of CHB. It is especially prevalent in that it was the only alternative therapy for patients who developed lamivudine resistance in the early 2000s. The lower rate of resistance and slower onset of resistance in ADV compared with lamivudine made it more ideal for long-term treatment of CHB at that time.17 Although newer and improved agents are now available, patients currently treated with 10 mg ADV with sustained viral suppression are generally kept on their current treatment.
Previous studies have shown that ADV nephrotoxicity is dose-related, and a great number of renal dysfunction incidences were reported. The initial study of 120 mg ADV for the treatment of human immunodeficiency virus reported a slow but steady increase in Cr, with 35% of patients experiencing renal dysfunction following 48 weeks of treatment and 50% of patients following 72 weeks of treatment.18 The dosage of ADV was reduced to 60 mg due to less frequent incidences of nephrotoxicity (27%) after 42 weeks of follow up.11, 19, 20 A study reported the incidence of renal dysfunction in 13% of patients treated with 30 mg ADV for 48 weeks.11 The 10-mg dose of ADV had a more favorable risk/benefit profile when compared with 30 mg ADV.17 A 5-year clinical trial of 10 mg ADV reported that four of 125 patients (3%) had confirmed increase in serum Cr of ≥0.5 mg/dL.3 We found that seven patients (5%) had an increase in serum Cr of ≥0.5 mg/dL in our cohort following a median follow-up of 30 months. A long-term study of lamivudine-resistant patients on 10 mg ADV and lamivudine combination therapy reported that 44 (16%) patients developed an increase in serum Cr of >0.5mg/dL over a 5-year period.21 Nevertheless, incidence of renal dysfunction may still be underreported, because previous studies have reported renal dysfunction based primarily on serum Cr levels.
We further analyzed the data using calculated eGFR for indication of renal dysfunction, because eGFR would better reflect renal function. We found a significant difference in the incidence and magnitude of decrease in eGFR in patients treated with ADV compared with a similar cohort unexposed to ADV. Similar trends were observed with the subanalysis using the MDRD formula, which takes into account serum Cr, age, sex, ethnicity, blood urea nitrogen concentration, and albumin concentration. This is likely due to the fact that none of our patients were African American, which would have had a significant impact on the calculation, and that the albumin values were largely within normal range for both the ADV group and the unexposed group (4.3 ± 0.5 g/dL and 4.4 ± 0.4 g/dL, respectively). We observed 10 (7%) patients with decrease in eGFR of >30% from baseline and 17 (11.7%) patients who developed eGFR ≤50 mg/mL during treatment with ADV. The median follow-up time was shorter for patients unexposed to ADV; however, when this was taken into account with incidence density calculation and survival analysis, there was still a significantly higher incidence of renal dysfunction in patients exposed to ADV.
Following adjustments for age, sex, presence of mild baseline renal insufficiency, and HTN and/or DM, 10 mg ADV was found to be a significant predictor for both significant decrease in eGFR (HR 2.3, P = 0.018) and the development of moderate renal dysfunction (HR 3.94, P = 0.030). The relative risk for development of moderately impaired renal function and/or treatment termination due to renal dysfunction was 3.68 in patients exposed to 10 mg ADV when compared with the unexposed group.
Physicians tend to monitor renal function through serum Cr level rather than eGFR. The current recommendation is to monitor serum Cr every 3 months in all patients who are predisposed to renal insufficiency and on ADV for more than 1 year.17 Clinicians managing patients with CHB treated with 10 mg ADV need to monitor patients' eGFR to observe potential renal dysfunction and adjust dose or terminate treatment accordingly.
Patients who were switched off ADV due to suspected and confirmed renal dysfunction were older and had lower baseline eGFR than other patients in our cohort. All patients who were switched off ADV recovered with their eGFR level returning to baseline. ADV is an independent predictor for significant deterioration of renal function. Patients on ADV should be monitored closely, especially those with baseline renal impairment, age >50, or HTN and/or DM.
There were limitations to our study. Patients were observed retrospectively. The unexposed group had a shorter duration of follow-up; however, the incidence density per 100 person-years was still significantly higher in the ADV group compared with the comparative unexposed group (5.0 cases versus 1.36 cases [P = 0.02]). The proportion of patients who maintain their eGFR within 20% from baseline was also significantly lower in the ADV group using the Kaplan-Meier method (P = 0.0104).
In conclusion, we observed greater incidences and severity of renal dysfunction as measured by decrease in eGFR in patients on 10 mg ADV when compared with patients in the unexposed group. Patients on ADV should be monitored closely for signs of renal dysfunction with eGFR measurements rather than serum Cr levels. Older patients, patients with renal insufficiency at baseline, and patients with HTN and/or DM are particularly at risk.