Renal insufficiency in human immunodeficiency virus (HIV) type 1 (HIV-1)-infected patients is becoming more common (
18). This is due in part to the development of an HIV-associated nephropathy (HIVAN), characterized by massive proteinuria and rapidly progressive renal failure, in a subset of patients infected with HIV (
8,
13,
24). Although combination antiretroviral therapy appears to slow the progression of HIVAN to end-stage renal disease (ESRD) in many cases (
9,
16,
27), the total number of persons with ESRD due to HIVAN is expected to increase due to the increasing prevalence of HIV-1 infection (E. J. Schwartz, L. Szczech, J. A. Winston, and P. E. Klotman, J. Am. Soc. Nephrol.
11:abstr. A0882, 2000). In addition to HIVAN, conditions such as hypertension, diabetes, or other glomerular diseases including membranoproliferative glomerulonephritis, amyloidosis, acute immune complex glomerulonephritis, cryoglobulinemia, and immunoglobulin A nephropathy may lead to ESRD in HIV-infected patients (
1,
4,
10,
14,
15,
19,
22-
25). HIV infection has become the third most frequent cause of renal failure among African Americans ages 20 to 64 (
18). As HIV-infected patients continue to live longer and as the number of HIV-infected patients with ESRD continues to increase, the importance of understanding the pharmacokinetic properties of antiretroviral drugs in patients with ESRD becomes more crucial.
Lamivudine (Epivir; GlaxoSmithKline) is an antiretroviral drug commonly used to treat HIV infection. It is a cytosine dideoxynucleotide analogue prodrug that is phosphorylated within cells to an active triphosphate form which acts as a potent inhibitor of HIV reverse transcriptase. Although active as monotherapy against HIV, lamivudine should always be used in combination with other antiretroviral agents in order to achieve more effective suppression of viral replication and prevent the development of resistance (
3,
5,
20,
21,
26). Antiretroviral combinations that include lamivudine decrease HIV viral loads to low or undetectable levels in a majority of patients (
5) and delay the development of AIDS-defining events or death (
6). Because it is primarily excreted in the urine, an understanding of the pharmacokinetic properties of lamivudine in patients with ESRD is important to guide dosing recommendations.
The pharmacokinetics of a single 300-mg dose of lamivudine were previously studied in HIV-infected subjects with normal renal function, moderate renal impairment, or severe renal impairment (
7). In comparison to subjects with normal renal function, those with impaired renal function had higher peak concentrations in serum, longer terminal elimination half-lives (
t1/2s), and larger areas under the serum concentration-time curves (AUCs). On the basis of the results of that study, the recommended dosage of lamivudine in patients with severe renal impairment is 25 mg once daily rather than the standard dose of 150 mg twice daily. However, the study did not evaluate the effect of dialysis on the pharmacokinetics of lamivudine. The low molecular weight, low level of protein binding, high degree of water solubility, and high degree of permeability exhibited by lamivudine suggest that it would be readily removed by dialysis (
11,
12). A single-dose pharmacokinetic study of lamivudine in HIV-negative subjects undergoing hemodialysis indicated that although hemodialysis removed up to 50% of dialyzed lamivudine, it did not reduce the concentrations in serum to a clinically significant extent because of a large apparent volume of distribution (
12). Neither of these single-dose studies assessed the pharmacokinetics of lamivudine at steady state.
The present study was designed to evaluate the pharmacokinetics of lamivudine at steady state in a cohort of HIV-infected subjects with ESRD receiving combination antiretroviral therapy that included lamivudine. All subjects were also undergoing chronic hemodialysis or chronic ambulatory peritoneal dialysis (CAPD).
MATERIALS AND METHODS
Patient population.
HIV-infected subjects with ESRD requiring dialysis (hemodialysis or CAPD) were recruited. Subjects were included in the study if they were men or nonpregnant nonlactating women ages 18 or older, had evidence of HIV-1 infection confirmed by a positive antibody test result, were receiving combination antiretroviral therapy that included lamivudine, and had no significant hematologic, hepatic, or pancreatic dysfunction. Subjects were not eligible to participate in the study if they had medical conditions that could compromise their safety or interfere with the pharmacokinetic measurements, were active substance abusers, or had a history of sensitivity to lamivudine. Nine subjects undergoing hemodialysis and two subjects undergoing CAPD were studied.
Study design.
This open-label, single-center study was approved by the Duke University Investigational Review Board, and all subjects gave written, informed consent before any study procedures were performed. All study participants had been receiving lamivudine for at least 3 months prior to enrollment in the study. At the time of enrollment, the lamivudine dosage was standardized to 150 mg orally once daily as part of combination antiretroviral therapy and was continued for at least 14 days with maintenance of the subject's usual dialysis schedule and other medications. For subjects receiving hemodialysis, lamivudine was taken shortly after the completion of dialysis on dialysis days. Subjects receiving hemodialysis were dialyzed three times per week; subjects receiving CAPD exchanged dialysate four times daily. After this 14-day lead-in period, subjects were admitted to the Duke Clinical Research Unit for sampling for pharmacokinetic studies.
For subjects receiving hemodialysis, a predose serum sample was obtained on hospital day 1. After oral administration of 150 mg of lamivudine, serum samples were obtained at 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 4, 6, 8, 10, 12, 16, and 20 h. Approximately 20 h after administration of the lamivudine dose hemodialysis was then performed over 3.5 to 4 h with GFS-20 dialysis membranes. A predialysis serum sample was obtained, as were simultaneous arterial and venous serum samples, at three time points during the dialysis session (early, midpoint, and late dialysis samples). Following hemodialysis, a second predose serum sample was drawn (corresponding to a sample obtained 24 h after dosing with dose 1) and a second 150-mg dose of lamivudine was administered. Serum samples were again obtained at 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 4, 6, 8, 10, 12, 16, 20, and 24 h postdosing.
For subjects undergoing CAPD, a predose serum sample was drawn on the morning of hospital day 1, a peritoneal dialysis exchange was completed, and a 150-mg dose of lamivudine was administered. Serum samples were then obtained at 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 4, 6, 8, 10, 12, 16, and 24 h postdosing. During this period, subjects continued their usual CAPD schedule. Immediately prior to each exchange, a sample of the peritoneal dialysate was collected. On hospital day 2, a predose serum sample was again drawn (corresponding to a sample obtained 24 h after dosing with dose 1) and a second 150-mg dose of lamivudine was administered. CAPD was withheld and serum samples were obtained by using the same sampling schedule described above for the next 24 h.
Assay of lamivudine levels.
All serum and dialysate samples were stored at −70°C and transferred on dry ice to Glaxo Wellcome, Inc., for analysis. Lamivudine concentrations in serum and peritoneal dialysate fluid were determined by a validated high-performance liquid chromatography (HPLC) assay, followed by tandem mass spectrometry by the multiple reaction monitoring (MRM) technique.
For serum samples, lamivudine and its stable isotopically labeled internal standard were extracted from serum by ultrafiltration with Amicon Centricon 30 concentrators. The filtrate was analyzed by HPLC with a reverse-phase column and detection by tandem mass spectrometry by the positive-ion MRM technique. The calibration range for this method is 2.5 to 5,000 ng/ml with 0.25 ml of serum or plasma. The concentrations of lamivudine in samples were determined by interpolation from calibration curves based on linear regressions of the ratios of the peak area for the lamivudine calibration standard to the peak area for the internal standard versus the corresponding nominal lamivudine concentrations by using a weighting factor of 1/concentration2. The interassay precision of the assay, expressed as the coefficient of variation, was less than 12.1%. The accuracy of the assay, expressed as percent bias, was less than 4.9%.
For dialysate samples, lamivudine and its stable isotopically labeled internal standard were extracted from dialysate solution by using Waters OASIS HLB 30-mg solid-phase extraction plates. The eluent was analyzed by HPLC with a reverse-phase column and detection by tandem mass spectrometry by the positive-ion MRM technique. The calibration range for this method is 2.5 to 600 ng/ml with a 0.10-ml dialysate solution. The concentrations of lamivudine in samples were determined by interpolation from calibration curves based on linear regressions of ratios of the peak area for the lamivudine calibration standard to the peak area for the internal standard versus the corresponding nominal lamivudine concentrations by using a weighting factor of 1/concentration2. The interassay precision of the assay, expressed as the coefficient of variation, was less than 9.1%. The accuracy of the assay, expressed as percent bias for validation control samples, was less than 12.4%.
Pharmacokinetic analysis.
Noncompartmental methods were used to determine the values of the pharmacokinetic parameters for lamivudine, including the AUC from time zero to 24 h (AUC0-24) and t1/2. The maximum concentration in serum (Cmax), the predose concentration in serum (C0), and the concentration in serum at 24 h postdosing (C24) were obtained by direct inspection of the time course data. These parameters were determined for each subject for each dose of lamivudine given. The results are represented as geometric means and 95% confidence intervals (CIs) for each parameter. Pharmacokinetic modeling was performed with WinNonlin software (version 2.1; Pharsight Corporation) to explore the feasibility of using different dosing regimens and schedules in our cohort undergoing hemodialysis. Data from the hemodialysis day were simulated by use of a one-compartment model with first-order absorption and elimination with an absorption lag (WinNonlin). Cmax, the time to Cmax, and AUC over the dosing interval were calculated for multiple dosing regimens, including 25 mg daily, 50 mg daily, 50 mg every other day, 75 mg every other day, and 150 mg weekly.
Simultaneous arterial and venous blood samples obtained during hemodialysis were used to calculate an extraction ratio, defined as (
Ca −
Cv)/
Ca, where
Ca is the concentration of lamivudine in arterial serum and
Cv is the concentration of lamivudine in venous serum. In addition, clearance of lamivudine by hemodialysis was calculated by the equation
QH · (1 −
H +
KH) · [(
Ca −
Cv)/
Ca], where
QH is the blood flow rate through the dialysis filter,
H is the hematocrit,
K is the concentration of lamivudine in red blood cells/concentration of lamivudine in plasma reported previously (
12), and
Ca and
Cv are as defined above. Since in vitro studies indicate that lamivudine partitions equally between red blood cells and plasma (
12), a value of 1 was used for
K. The amount of lamivudine removed during hemodialysis was estimated by the following formula: [lamivudine] · ER ·
F ·
T, where [lamivudine] is the mean arterial lamivudine concentration during the dialysis session, ER is the extraction ratio,
F is the blood flow rate, and
T is the duration of dialysis. The amount of lamivudine removed by CAPD was determined by multiplying the concentration of lamivudine in the dialysate at the end of each dialysis dwell by the dialysate volume.
Safety.
Adverse events were monitored by history taking, physical examination, laboratory tests, and electrocardiogram at the time of enrollment. History taking, physical examination, and laboratory tests were also performed during the hospitalization, and patients were questioned periodically throughout the study regarding possible adverse effects.
RESULTS
Nine subjects receiving hemodialysis and two subjects receiving CAPD successfully completed the study with no treatment-related adverse events. All study subjects were male, 10 were African American, and 1 was Caucasian. Subjects had a median age of 49 years (range, 38 to 56 years), a median weight of 67 kg (range, 56 to 88 kg), a median CD4+ absolute T-cell count of 338 cells/mm3 (range, 51 to 634 cells/mm3), and a median CD4+ percentage of 19% (range, 7 to 35%). More than half of the subjects had HIV RNA levels below the lower limit of detection; thus, the median was below the lower limit of detection of <400 copies/ml. The range of HIV RNA levels was <50 to 71,770 copies/ml. Subjects had been diagnosed with HIV infection for a median of 5 years (range, 1 to 11 years), had been dialyzed for a median of 19 months (range, 3 to 71 months), and had been receiving lamivudine therapy for a median of 22 months (range, 5 to 48 months). The usual lamivudine dosing regimen prior to participation in the study was 150 mg orally once daily in seven subjects, 150 mg orally three times a week after dialysis in three subjects, and 150 mg orally twice daily in one subject.
Nondialysis day.
Steady-state pharmacokinetics were similar on the nondialysis day for subjects undergoing hemodialysis and subjects undergoing CAPD and are depicted in Tables
1 and
2, respectively. In the cohort undergoing hemodialysis, the observed geometric mean
Cmax was 3.24 μg/ml (95% CI, 2.61 to 4.02 μg/ml) and the observed geometric mean AUC
0-24 was 46.5 μg · h/ml (95% CI, 37 to 58.4 μg · h/ml). Serum lamivudine concentrations before dosing and at 24 h postdosing were similar: 0.81 μg/ml (95% CI, 0.59 to 1.1 μg/ml) and 1.28 μg/ml (95% CI, 0.96 to 1.71 μg/ml), respectively. The apparent
t1/2 was 17.2 h (95% CI, 10.5 to 28.1 h). Among the cohort undergoing CAPD, the observed geometric mean
Cmax was 4.5 μg/ml and the observed geometric mean AUC
0-24 was 65.7 μg · h/ml. Serum lamivudine concentrations before dosing and at 24 h postdosing were similar: 1.77 and 1.88 μg/ml, respectively. The apparent
t1/2 among these subjects was 20 h.
Dialysis day.
Pharmacokinetic parameters observed under the influence of dialysis were similar to those observed on a nondialysis day both for subjects undergoing hemodialysis and for subjects undergoing CAPD (Tables
1 and
2), except that the apparent intradialysis
t1/2 among subjects undergoing hemodialysis was reduced to 5.3 h (Table
1). A graphical representation of the mean serum lamivudine concentration-versus-time profile for the hemodialysis cohort is shown in Fig.
1A. Comparison of simultaneous concentrations in arterial and venous blood during the hemodialysis session demonstrated that the lamivudine extraction ratios remained fairly constant throughout the session, with a mean extraction ratio of 0.318, verifying the removal of lamivudine during hemodialysis. The geometric mean clearance of lamivudine during hemodialysis among the nine study subjects was calculated to be 138 ml/min (range, 61 to 191 ml/min), and the amount of lamivudine removed during the dialysis session was estimated to be approximately 28 mg. Interestingly, mean serum lamivudine levels rose rapidly immediately after the hemodialysis session, prior to administration of the second dose of lamivudine, suggesting that lamivudine was redistributed into the serum, probably from the intracellular compartment. Thus, although hemodialysis extracted a significant amount of lamivudine from serum, the elimination of lamivudine by dialysis was not of sufficient magnitude to cause a statistically significant change in
Cmax or AUC
0-24.
Pharmacokinetic modeling of our data indicated that a lamivudine dose of 25 mg once daily among HIV-infected patients undergoing hemodialysis would provide concentrations in serum approximating those seen in patients with normal renal function receiving 150 mg twice daily (Fig.
2) (
2). In our subject with the lowest AUC for the 24-h dosing interval, a postdialysis dose of 25 mg daily would yield an AUC of 4.42 μg · h/ml, with a
Cmax of 206 ng/ml and a
C0 of 100 ng/ml.
Among the subjects undergoing CAPD the amount of lamivudine removed by peritoneal dialysis was calculated by measuring the amount of lamivudine in the dialysate. The total measured amount of lamivudine removed in the peritoneal dialysate over a 24-h period was 23.5 mg in one subject and 24.8 mg in the other. For these subjects, pharmacokinetic profiles for lamivudine were obtained in the presence of CAPD (dose 1) or the absence of CAPD (dose 2) (Table
2). Figure
1B shows a graphical representation of the mean concentration-time profile for these two subjects. Although CAPD removed approximately 16% of the daily oral dose of lamivudine, withholding CAPD for 1 day did not significantly alter the pharmacokinetic parameters.
DISCUSSION
The purpose of this study was to evaluate the steady-state pharmacokinetics of lamivudine in HIV-infected patients with ESRD in order to determine the appropriate dose for such subjects. We believed that it was important to perform our evaluations at steady state because single-dose studies may fail to recognize the contributions of the drug distribution in tissue or alternative routes of metabolism or elimination induced by prolonged exposure to a drug. For example, in addition to renal elimination, lamivudine is also metabolized through a hepatic pathway (
12). Although in patients with normal renal function hepatic metabolism contributes little to the overall metabolism of the drug, this pathway becomes more important in patients with renal impairment (
12). In clinical practice, a dose of 150 mg of lamivudine given orally one daily is well tolerated and has commonly been used for patients with ESRD undergoing hemodialysis. The data from this study suggest, however, that this dose is considerably higher than the dose needed. In HIV-infected patients with normal renal function taking lamivudine at a dose of 150 mg orally every 12 h, the AUC
0-12 was 4.57 to 6.63 μg · h/ml (
2), whereas the AUC over a 24-h dosing interval was 49.8 μg · h/ml in our hemodialysis subjects taking 150 mg of lamivudine orally once daily. Pharmacokinetic modeling of data for the subject with the lowest AUC in our hemodialysis cohort indicated that administration of a dose of 25 mg once daily would yield an AUC over the dosing interval approximating that for subjects with normal renal function receiving 150 mg twice daily. Because of the prolonged
t1/2 of lamivudine in patients with ESRD, an alternative dosing regimen of 75 mg every other day would be expected to provide similar levels (Fig.
2). For hemodialysis patients for whom adherence to medication regimens is an issue, this could potentially provide a postdialysis dosing schedule that would allow directly observed therapy. This would be most helpful if other antiretroviral agents could be administered in a similar fashion.
These alternative lamivudine dosing regimens used in the present study (25 mg daily or 75 mg every other day) were suggested on the basis of pharmacokinetic data derived from measurements obtained with serum. Potentially, serum lamivudine levels do not accurately reflect the level of the active triphosphate form of lamivudine at the site of action in HIV-infected cells. Since lamivudine triphosphate concentrations in peripheral blood mononuclear cells have been measured and correlated with the concentrations in serum, reasonable inferences regarding intracellular concentrations in our cohort can be made (
17). Recognizing, however, the dangers inherent in underdosing and the excellent tolerability of the higher lamivudine doses used in this study, the common practice of dosing patients with ESRD with doses higher than 25 mg remains a reasonable option.
Although our study was not designed to evaluate differences between hemodialysis and CAPD on the pharmacokinetics of lamivudine, our results indicate that the amount of lamivudine removed by dialysis and the values of the pharmacokinetic parameters such as AUC0-24 were similar in both groups. Hemodialysis removed approximately 28 mg of lamivudine during a standard 3.5-h session. This was sufficient to change the apparent lamivudine t1/2 from approximately 16 h in the absence of hemodialysis to 5.2 h during the dialysis session. Immediately after hemodialysis, the serum lamivudine level increased, probably due to redistribution from the intracellular compartment. One implication of this finding is that hemodialysis may not be effective in eliminating lamivudine from the body in the setting of a lamivudine overdose. Our data also demonstrated that a measurable amount of lamivudine, 16% of the daily dose administered, was eliminated from the body in the peritoneal dialysate in the patients undergoing CAPD. Although both hemodialysis and CAPD removed lamivudine, the magnitude of the elimination by either route was insufficient to meaningfully alter the measured AUC0-24 of the drug. This suggests that the amount of lamivudine removed by dialysis (CAPD or hemodialysis) is small relative to the total body pool of lamivudine.
Finally, none of our study participants experienced adverse events related to lamivudine therapy, even though many of these patients had been taking lamivudine at a dose of 150 mg daily for many months and had AUC0-24s five times those for patients with normal renal function. This suggests that lamivudine is a relatively safe drug with a high therapeutic index. Given the increasing prevalence of HIV-infected patients with ESRD, understanding the pharmacokinetics of lamivudine is important to enable clinicians to treat their patients safely and effectively.
Acknowledgments
This study was supported by research grants from Glaxo Wellcome Research and Development and from NIAID, NIH (grant 1K24AI01608-01). This study was performed in the Duke Clinical Research Unit, which is supported by grant MO1-RR-30 from the National Center for Research Outcomes, Clinical Research Centers Program, NIH.
We are grateful for the contributions of Ken Shipp, who is a clinical pharmacist, and Jason Stout of the Division of Infectious Diseases, Duke University Medical Center. We also acknowledge Glaxo Wellcome, Inc., and Laurel M. Adams for performing the assays for lamivudine with our samples. We also acknowledge the support of the Centers for AIDS Research at Duke University and the University of North Carolina—Chapel Hill.