Pharmacokinetics and safety of oseltamivir in patients with end-stage renal disease treated with automated peritoneal dialysis

Abstract

AIMS

Patients with end-stage renal disease (ESRD) are at increased risk of developing complications associated with influenza infection. Oseltamivir is indicated for influenza treatment in ESRD patients, but the disposition is poorly understood in this patient population. This study aimed to characterize the pharmacokinetics and tolerability of oseltamivir in automated peritoneal dialysis (APD) and construct a pharmacokinetic model to assist with optimized dosing.

METHODS

Ten adults with ESRD were prescribed an aggressive APD regimen consisting of three continuous cycler-assisted peritoneal dialysis (CCPD) sessions during the day and two continuous ambulatory (CAPD) sessions overnight. Oseltamivir was administered as a single 75 mg dose, immediately before APD treatment.

RESULTS

Oseltamivir was rapidly eliminated via first-pass metabolism, with most of the dose (Fraction metabolized = 0.964) reaching the circulation as the active metabolite, oseltamivir carboxylate. This metabolite was cleared slowly and was quantifiable throughout the sampling interval. The disposition of oseltamivir and oseltamivir carboxylate was described by a two- and a one-compartment model, respectively. Metabolite clearance by CCPD [0.32 l h⁻¹ (70 kg)⁻¹] was 1.9-fold faster than via CAPD [0.17 l h⁻¹ (70 kg)⁻¹], with renal elimination being dominant in patients with residual urine production. Model simulations showed that a single 75 mg dose attained target exposures in patients with negligible or low urine clearance. However, higher doses are recommended for further investigation in patients with high residual renal function. In all patients, oseltamivir was well tolerated.

CONCLUSIONS

In APD patients with anuria or low residual renal elimination, a single 75 mg dose of oseltamivir produced exposures at the upper end of the safety margin.

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT

• The disposition of oseltamivir in patients with end-stage renal disease undergoing automated peritoneal dialysis (APD) is poorly understood.

• Drug clearance in APD is higher than in traditional continuous ambulatory peritoneal dialysis.

WHAT THIS STUDY ADDS

• The disposition and tolerability of a single 75 mg dose was assessed in patients on an aggressive simulated APD regimen.

• In patients with negligible or low residual renal elimination, a 75 mg dose of oseltamivir produced exposures at the upper end of established safety ranges and was generally well tolerated.

• This study provides new information regarding oseltamivir dosing in patients receiving APD treatment. It provides nephrologists with clearer guidelines on the APD and drug dosing regimens required to achieve optimal therapy in this patient cohort.

Introduction

Peritoneal dialysis (PD) is a form of renal replacement therapy that is widely used in patients with end-stage renal disease (ESRD). It involves the infusion of large volumes of dialysis solution into the peritoneal cavity, thereby allowing for the exchange of fluids and solutes across the peritoneum. The use of automated forms of PD has increased in recent years [1], primarily because traditional continuous ambulatory PD (CAPD) relies on the manual delivery and drainage of dialysate. In contrast, automated PD (APD) employs a mechanical device known as a cycler to facilitate dialysate exchanges. These automated exchanges typically occur overnight while the patient is asleep, resulting in minimal disruption to the patient's daily routine. Furthermore, APD has been associated with lower rates of peritonitis, fewer mechanical complications and the potential for increased small solute clearance [2,3].

Commonly used APD modalities are illustrated in Supplementary Figure S1. In continuous cycler-assisted PD (CCPD), overnight exchanges are followed by one or two long daytime exchanges, while nightly intermittent PD (NIPD) involves APD at night only (‘dry day’). Combination therapy of CCPD or NIPD with CAPD is possible (Supplementary Figure S1). Generally, PD prescriptions are highly customized to the individual patient and take into account the body size, residual renal function and peritoneal permeability [4]. To ensure the attainment of target exposures, the clinician must also consider dialysis-specific factors, such as fill volumes, the number of daily exchanges and the mode of PD [5]. Aggressive dialysis is achieved using large fill volumes and frequent exchanges, both of which are feasible with APD (up to 25 l day⁻¹). However, rapid cycling with high volumes of hypertonic glucose solution may cause sodium sieving and poor sodium removal that results in hypernatraemia [5]. The choice of osmotic agent is also important; icodextrin-based solutions offer the advantages of sustained ultrafiltration and increased small solute clearances during long-dwell exchanges [6].

Patients with ESRD are among those at elevated risk of developing complications if they are infected with influenza [7,8]. While vaccination is recommended as the primary intervention for high-risk patients [8], the neuraminidase inhibitor oseltamivir (Tamiflu®) is indicated for the treatment and prophylaxis of influenza A and B infections [9,10].

Oseltamivir is an orally administered phosphate prodrug that is rapidly converted to the active metabolite, oseltamivir carboxylate. The key pharmacokinetic properties of oseltamivir and oseltamivir carboxylate were first established in the late 1990s [11]. Since then, researchers have gained an understanding of the disposition of both compounds in specific populations, including patients with ESRD undergoing haemodialysis or CAPD [12]. However, data are lacking in patients receiving APD, in whom exposures may differ from subjects treated with CAPD due to the potential for higher drug clearance [13]. This information is necessary to allow for the design of optimized dosing regimens in patients undergoing both APD and oseltamivir treatment.

The present study (NCT01556633) characterized the pharmacokinetics of oseltamivir and oseltamivir carboxylate in adult ESRD patients receiving a rapid cycling dialysis regimen designed to simulate an aggressive APD prescription. Safety and tolerability were also assessed. In addition, the data were used to develop a population pharmacokinetic model for oseltamivir that can assist with the selection of optimal dosing strategies in these patients.

Methods

Study design and conduct

This was a prospective, open-label, single oral dose phase I study conducted at two specialist clinical study facilities in New Zealand. All patients provided written, informed consent, and the study was carried out in accordance with the requirements of the Declaration of Helsinki and local regulations. The study protocol was approved by the relevant institutional review boards or ethics committees at the participating study centres. The enrolment target was 10 subjects. The study was approved by the Upper South Island Regional ethics committee (URB/12/02/002). The study is registered at ClinicalTrials.gov (ID number NCT01556633).

Patients were aged 19–90 years and were required to have the following characteristics: receiving stable dialysis therapy for ≥3 months prior to study day 1; body surface area 1.7–2.3 m²; a total Kt/V (where K is the dialysis clearance of urea, t the dialysis time and V the volume of distribution of urea) >1.7 [14]; a peritoneal equilibration test result that did not indicate low transporter status in the 3 months preceding study day 1.

Subjects were excluded for the following reasons: they were pregnant or lactating; had clinically significant or unstable disease not linked to renal impairment or a recent clinical history considered significant by the investigator; had uncontrolled systolic or diastolic blood pressure; had resting bradycardia (<45 beats min⁻¹) or tachycardia (mean rate >100 beats min⁻¹); had the need or potential need for significant change of medical management during the study; had decompensated liver disease or presence of hepatitis B or C or human immunodeficiency virus antibodies; showed presence of clinically significant laboratory abnormalities; had a history of drug or alcohol abuse or of smoking ≥10 cigarettes day⁻¹ or tobacco equivalent; or were taking medication that might affect serum creatinine concentrations or that might interfere with renal excretion of drugs in the 2 weeks preceding the study.

Treatments

Patients were converted to a standardized, broadly applicable APD regimen comprising three CCPD exchanges (3 × 2.5 litres) over an 8 h period during the day and two CAPD sessions (2 × 2.0 litres) over 16 h overnight (Supplementary Figure S1). This regimen had previously been used successfully [15] and is in alignment with The National Kidney Foundation guidelines for effective peritoneal dialysis [14]. Dextrose (1.5–4.25% w/v) was used for all exchanges over 24 h except for one 8 h CAPD exchange, in which icodextrin 7.5% w/v was the osmotic agent. Patients were transferred to this rapid cycling regimen from their usual prescription in order to mimic conditions of a ‘worst-case’ low drug exposure scenario; such scenario is a consequence of rapid drug clearance that produces inadequate antiviral efficacy and elevated risk of resistance. Evidence suggests that the higher exchange volumes associated with aggressive dialysis regimens (including rapid cycling with frequent exchanges) might increase systemic clearance of some drugs [13]. To account for this, the present PD regimen, which involves rapid cycling and two CAPD exchanges, was designed with the aim of achieving high drug clearance. The effectiveness of this regimen was further enhanced by the use of icodextrin during one of the CAPD sessions.

Oseltamivir was administered as a single 75 mg oral dose on day 1, with APD commencing immediately afterwards to simulate high rates of drug clearance associated with rapid cycling. As the dialysis included a significant cycler-assisted component, a 75 mg dose was selected empirically to reflect concerns that the recommended dose in CAPD (30 mg) may not achieve critical concentrations at 120 and 168 h postdosing. This has been observed with fluconazole, which is cleared twice as rapidly with APD than with CAPD and has a similar molecular weight, protein binding and volume of distribution to oseltamivir [13].

Standardized meals were provided on day 1 and at least 30 min before receiving the study medication. Fluids were consumed with meals and at a rate of ≥200 ml every 4 h from day 1 to day 8. Foods and beverages containing caffeine, alcohol or grapefruit juice were not provided. Smoking was not permitted during the 2 h before and 2 h after study drug administration. Light ambulatory activities were allowed.

Study assessments

Screening occurred from day −28 to day −2; baseline measurements were obtained on day −1.

Pharmacokinetics

Patients remained in the unit until day 8, when the last pharmacokinetic samples were collected. Follow-up occurred from day 15 to day 22. Plasma samples were obtained for the determination of oseltamivir and oseltamivir carboxylate concentrations prior to dosing and at 0.5, 1.33, 2, 2.5, 3, 4, 5, 6.67, 8, 10, 12, 14, 16, 20, 24, 28, 32, 48, 72, 96, 120, 144 and 168 h postdose. Dialysate was collected at the beginning and end of each PD interval for the first 48 h and then every 24 h. Urine was obtained prior to study drug administration and at intervals of 0–24, 24–48, 48–72, 72–96, 96–120, 120–144 and 144–168 h postdosing. The intensive plasma sampling was designed to allow for adequate estimation of the clearance by dialysis components, hence the area under the concentration–time curve extrapolated to infinity (AUC_0–inf). This collection schedule also allowed the individual estimation of oseltamivir clearance by CCPD or CAPD regimen. Overall, the study design was intended to derive broadly applicable dosing regimens that would account for the clearances for a range of APD prescriptions.

Noncompartmental analysis was conducted to guide initial parameterization of the population pharmacokinetic model. The following variables were assessed for both oseltamivir and oseltamivir carboxylate: total dialysate clearance (CLD_APD); clearance attributable to CCPD (CLD_CCPD) and CAPD (CLD_CAPD); peak plasma concentration (C_max), time to C_max (t_max), plasma elimination half-life (t_1/2), areas under the curve of plasma concentration vs. time to 120 and 168 h and extrapolated to infinity (AUC_0–120, AUC_0–168 and AUC_inf), and plasma concentrations at 120 and 168 h and at the time of the last measurable concentration (C₁₂₀, C₁₆₈ and C_last). Urinary parameters included renal clearance (CL_R) and cumulative drug excretion (Ae_0–t). The 120 and 168 h time points were selected to reflect the cumulative exposures following 5 day (treatment) and 7 day (prophylaxis) dosing with an oral 75 mg twice daily regimen.

The measurement of oseltamivir and oseltamivir carboxylate in plasma, dialysate and urine was performed using a previously validated liquid chromatography–mass spectrometry assay [16], which was sufficiently sensitive, accurate and precise. The plasma assay method had lower limits of quantification (LOQ) of 1.00 and 10.0 ng ml⁻¹ for oseltamivir prodrug and oseltamivir carboxylate, respectively; overall coefficients of variation (CVs) ranged from 2.7 to 7.6% and from 2.6 to 3.8%; and overall bias ranged from 3.0 to 0.3% and from 2.3 to 1.8%. The dialysate assay method had an LOQ of 0.500 and 5.00 ng ml⁻¹ for oseltamivir and oseltamivir carboxylate, respectively; overall CVs ranged from 2.4 to 5.1% and from 2.2 to 5.5%; and overall bias ranged from −0.5 to 1.3% and from 0.7 to 0.8%. The urine assay method had an LOQ of 5.00 and 30.0 ng ml⁻¹ for oseltamivir and oseltamivir carboxylate, respectively; overall CVs ranged from 2.3 to 5.8% and from 1.5% to 3.0%; and overall bias ranged from −4.0 to 0.0% and from 1.2 to 1.1%.

Pharmacokinetic modelling

The time course of oseltamivir and oseltamivir carboxylate in plasma, dialysate and urine was analysed by nonlinear mixed-effects modelling (NONMEM version 7.2; ICON Development Solutions, Hanover, MD, USA). Data analysis was performed using the first-order conditional estimation method with interaction and ADVAN9 to solve the differential equations. Models were selected using diagnostic scatter plots, the biological plausibility of parameter estimates and the objective function value (OBJ). A χ² test was used for comparison of nested models, where a decrease in the OBJ of 3.84 units (α < 0.05) was defined as statistically significant. Values below the LOQ were handled using the Beal M3 method [17].

Initially, models were developed that simultaneously described the concentration–time profiles of oseltamivir and oseltamivir carboxylate in plasma. These were then modified to include dialysate and urine data, for the explicit estimation of drug clearance via CCPD, CAPD and renal routes of elimination. Dialysate and urine compartments were ‘emptied and reset’ in the NONMEM data file after each collection period. For oseltamivir carboxylate, the central volume of distribution (Vc_OC) was fixed to that previously established {32.8 l (70 kg)⁻¹; Rayner et al. [18]} to avoid issues with parameter identification. The structure of the final model is illustrated in Figure 1, and differential equations are provided in the Supplementary Methods.

Figure 1.

Structure of the model describing the pharmacokinetics of oseltamivir and oseltamivir carboxylate in patients receiving oral oseltamivir and automated peritoneal dialysis. See Table 2 for parameter definitions

The between-subject variability (BSV) was estimated using an exponential variability model, with evaluation of various covariance–variance matrices. Residual unidentified variability was assessed using an additive, exponential or combined error model for each dependent variable (i.e. oseltamivir or oseltamivir carboxylate in plasma, dialysate and urine). The final model was evaluated by goodness-of-fit plots and by performing a visual predictive check (VPC). The VPC simulated 1000 data sets using the final parameter estimates, and is presented by comparison of the predicted median and 90% percentile range against corresponding observed data.

Monte Carlo simulations

Monte Carlo simulations were performed using NONMEM. The simulations included single oral doses of 30, 45, 60 and 75 mg (current regimen), a 75 mg dose administered at 0 and 48 h, or a dose of 30 mg daily over five consecutive days (recommended with severe renal impairment). For each dosing schedule, Monte Carlo simulations generated concentration–time profiles and AUC predictions. The AUC and concentration of oseltamivir carboxylate achieved at 120 h postdose were compared with mean historical data in healthy adult volunteers (AUC_0–120 = 31820 ng h ml⁻¹ and C₁₂₀ = 170 ng ml⁻¹; Rayner et al. [18]). These historical end-points were defined from previous studies that compared the relationship between oseltamivir exposures and viral shedding time. The duration of viral shedding is a key factor in determining the rate and extent of influenza infection. Previous studies have demonstrated that viral shedding time is significantly reduced in subjects who attain the defined oseltamivir target exposures, relative to placebo [19]. For each dosing regimen, the probability of target attainment (PTA) was calculated as the ratio of patients achieving the defined AUC_0–120 or C₁₂₀ target exposures.

Three APD regimens were used to simulate oseltamivir carboxylate exposures from the final model. These were as follows: (i) a standard CAPD-only regimen comprising three exchanges over 15 h during the day, followed by a single 9 h session overnight; (ii) an intermediate APD consisting of a CAPD exchange over 10 h before a 2 h break during the day, then three CCPD sessions over 9 h followed by a 3 h break overnight; and (iii) the intensive APD regimen used in the present study. A schematic representation of the three simulated APD prescriptions is shown in Supplementary Figure S1. The PTA for AUC_0–120 or C₁₂₀ of oseltamivir carboxylate was then calculated against historical standards as described above.

Safety

Adverse events were recorded by the investigator and coded according to MedDRA Version 14.1. Treatment-emergent adverse events (TEAEs) were those that commenced or worsened on or after study drug administration. Events were presented by system organ class and preferred term. Adverse events were graded as mild (no disruption of normal activities), moderate (sufficient to reduce or affect daily activity) or severe (inability to work or perform normal activities). The relationship to treatment was assessed by the investigator.

Serious adverse events were those that were fatal or life threatening, required prolonged hospitalization, caused significant or persistent incapacity or disability, or were congenital or birth abnormalities. Clinically significant laboratory abnormalities were those accompanied by clinical symptoms or those that required a change in therapy. Vital signs (blood pressure, heart rate and temperature) were recorded prior to the collection of pharmacokinetic blood samples; 12-lead electrocardiograms were recorded in triplicate (PR, QRS and QT intervals; QTcF, QTcB, pulse rate, U wave and T wave; overall assessment).

Results

Study population

Of the 27 patients screened, 10 were enrolled into the primary study cohort. All 10 were included in the safety evaluation; one patient was excluded from the pharmacokinetic assessment due to a protocol violation (QTcF found to be outside the permitted range after administration of study drug). Key baseline characteristics and demographics are shown in Table 1. The mean (range) age was 52.8 (24–70) years, and 50% of patients were male. Five patients were anuric and five had residual urine production. All subjects were using concomitant medication.

Table 1.

Individual patient demographics and estimates of oseltamivir carboxylate clearance by automated peritoneal dialysis or urine elimination

Patient	Anuric	Geographical ancestry	Age (years)	Bodyweight (kg)	Gender	Height (cm)	BSA (m2)	SCr (μmol l−1)	APD clearance* [l h−1 (70 kg)−1]	Urinary clearance [l h−1 (70 kg)−1]
1	No	Caucasian	70	89	Female	170	2.06	831	0.395	0.195
2†	No	Maori	64	92	Male	170	2.09	980	–	–
3	No	Pacific Islander	65	85	Male	182	2.07	698	0.356	0.382
4	Yes	Caucasian	24	69	Male	177	1.84	1461	0.366	–
5	Yes	Pacific Islander	48	60	Female	166	1.66	1307	0.269	–
6	No	Indian/African	64	79	Male	163	1.90	346	0.348	2.43
7	Yes	Indian/African	58	87	Male	168	2.02	1043	0.366	–
8	Yes	Indian/African	47	71	Female	161	1.78	1004	0.301	–
9	Yes	Pacific Islander	63	81	Female	163	1.91	729	0.376	–
10	No	Pacific Islander	25	76	Female	164	1.86	320	0.319	2.58

Abbreviations are as follows: APD, automated peritoneal dialysis; BSA, body surface area; CAPD, continuous ambulatory peritoneal dialysis; CCPD, continuous cycler-assisted peritoneal dialysis; SCr, serum creatinine concentration.

^*Estimates for elimination via APD are the sum of individual clearances by CCPD and CAPD.

^†Patient excluded from data analysis due to violation of study protocol.

Pharmacokinetics

The mean and individual pharmacokinetic profiles in plasma are presented in Figure 2. Noncompartmental parameters are summarized in Supplementary Table S1. Oseltamivir was rapidly removed from the plasma via first-pass metabolism (below LOQ within 10 h; Figure 2A), with most of the dose reaching the circulation as oseltamivir carboxylate. Less than 1% of the 75 mg dose was recovered unchanged in urine and dialysate. Oseltamivir carboxylate was eliminated slowly and was quantifiable throughout the 168 h sampling interval (Figure 2B). The noncompartmental geometric mean t_1/2 was 28.7 h; the AUC_0–inf, AUC_0–120 and AUC_0–168 of oseltamivir carboxylate were 93 800, 83 400and 89 200 ng h ml⁻¹, respectively. These exposures were higher than those reported in patients undergoing CAPD and in subjects with normal renal function treated with oral oseltamivir 75 mg twice daily orally (see Discussion) [12]. The geometric mean C_max was attained at t_max 20 h.

Figure 2.

Arithmetic mean (±SD) plasma concentrations of oseltamivir (A) and oseltamivir carboxylate (B) on a logarithmic scale vs. time after a single 75 mg oral dose in patients with end-stage renal disease undergoing dialysis. Spaghetti plots of individual patient traces for oseltamivir (C) and oseltamivir carboxylate (D) are also provided. , Subject 10001; , Subject 10003; , Subject 10004; , Subject 10005; , Subject 10006; , Subject 10007; , Subject 10008; , Subject 10009; , Subject 10010

Urine output was generally <1000 ml day⁻¹, with the exception of a single patient, in whom urine output ranged from 1463 to 2776 ml day⁻¹; this subject had a baseline creatinine clearance of <6 ml min⁻¹ and a CL_R of 0.331 l h⁻¹. For two other urine-producing patients with baseline creatinine clearance values of 13 and 14 ml min⁻¹, respective CL_R estimates were 2.17 and 1.94 l h⁻¹. These two subjects had dialysate and plasma concentrations of oseltamivir carboxylate that were below the LOQ after 72 and 96 h, respectively. For all other patients, oseltamivir carboxylate was detected in the dialysate throughout the entire collection period. The geometric mean for CAPD clearance of oseltamivir carboxylate using icodextrin was 0.195 l h⁻¹ and was comparable to that (0.177 l h⁻¹) by dextrose exchange. Hence, icodextrin had minimal impact on oseltamivir pharmacokinetics.

Pharmacokinetic modelling

Initial visual inspection of the pharmacokinetic data identified a considerable delay (∼28 h) between the maximal concentration of parent drug and oseltamivir carboxylate. To account for this delay, a hypothetical transit compartment was incorporated in the model and represented the known conversion of oseltamivir to oseltamivir carboxylate via hepatic first-pass metabolism (see Figure 1). A substantial fraction of the oseltamivir dose (F_met = 0.964) was estimated for first-pass conversion to metabolite, with the remainder absorbed as unchanged prodrug into the central compartment. Exclusion of this ‘first-pass’ compartment resulted in a significantly worse model (ΔOBJ = +397) that did not fit the input of oseltamivir carboxylate in plasma. The first-order rate constant for metabolite absorption (ka_OC = 0.109 h⁻¹) was approximately 12 times slower than estimated for unchanged oseltamivir (ka_OP = 1.36 h⁻¹).

The plasma disposition of oseltamivir was best described by a two-compartment model. Measured amounts of prodrug in dialysate and urine were negligible and did not support the clearance of unchanged oseltamivir via these routes. However, inclusion of a clearance term describing the plasma conversion of oseltamivir to oseltamivir carboxylate [CL_pm = 9.63 l h⁻¹ (70 kg)⁻¹] did produce a minor improvement in the model (ΔOBJ = −9.68).

A one-compartment model described the disposition of oseltamivir carboxylate in plasma. For this model, a previously established volume of distribution was assumed [Vc_OC fixed at 32.8 l (70 kg)⁻¹] with a relatively low between-subject variability (BSV = 38.7%) term estimated for this parameter. The clearance of oseltamivir carboxylate by CCPD [CL_OCCCPD = 0.319 l h⁻¹ (70 kg)⁻¹] was almost twice that by CAPD [CL_OCCAPD = 0.170 l h⁻¹ (70 kg)⁻¹]. In patients with residual urine production (n = 4), renal elimination was the major route of metabolite clearance [CL_OCURINE = 0.736 l h⁻¹ (70 kg)⁻¹], although a high BSV (117%) was calculated. The model also required a clearance term to represent other (unidentified) routes of oseltamivir carboxylate metabolism or elimination. A summary of parameter estimates for the final model is provided in Table 2.

Table 2.

Population pharmacokinetic parameter estimates for oseltamivir and oseltamivir carboxylate following oral dosing of oseltamivir to patients receiving automated peritoneal dialysis

Kinetic model	Parameter	Description of parameter	Units	Mean estimate	BSV (CV%)
Absorption	Fmet	Fractional conversion to metabolite via first pass	–	0.964	16.9
	kaOP	Rate constant for oseltamivir absorption	1 h−1	1.36	147
	AlagOP	Lag time in absorption of oseltamivir	h	0.485	53.8
	kaOC	Rate constant for oseltamivir carboxylate absorption	1 h−1	0.109	–
Oseltamivir disposition	CLpm	Clearance of oseltamivir to metabolite in plasma	l h−1 (70 kg)−1	9.63	–
	VcOP	Central distribution volume of oseltamivir	l (70 kg)−1	16.7	31.1
	CLD	Intercompartmental clearance of oseltamivir	l h−1 (70 kg)−1	6.80	–
	VpOP	Peripheral distribution volume of oseltamivir	l (70 kg)−1	307	–
Oseltamivir carboxylate disposition	VcOC	Central distribution volume of oseltamivir carboxylate	l (70 kg)−1	32.8 FIX*	38.7
	CLOCCCPD	Clearance of oseltamivir carboxylate by CCPD regimen	l h−1 (70 kg)−1	0.319	7.70
	CLOCCAPD	Clearance of oseltamivir carboxylate by CAPD regimen	l h−1 (70 kg)−1	0.170	7.70
	CLOCURINE	Clearance of oseltamivir carboxylate by renal elimination	l h−1 (70 kg)−1	0.736	117
	CLOCOTH	Clearance of oseltamivir carboxylate by other routes	l h−1 (70 kg)−1	0.319	36.9
Residual error†	CVOP_plasma	Proportional error for oseltamivir in plasma	CV%	–	35.9
	SDOC_plasma	Additive error for oseltamivir carboxylate in plasma	μg l−1	–	73.1
	CVOC_dial	Proportional error for oseltamivir carboxylate in dialysate	CV%	–	14.0
	SDOC_dial	Additive error for oseltamivir carboxylate in dialysate	μg	–	11.6
	CVOC_urine	Proportional error for oseltamivir carboxylate in urine	CV%	–	35.5

Abbreviations are as follows: BSV, between-subject variability; CAPD, continuous ambulatory peritoneal dialysis; CCPD, continuous cycler-assisted peritoneal dialysis; CV, coefficient of variation.

^*Central compartment volume of distribution was fixed to that previously established to avoid parameter identifiability (see Methods).

^†Data were modelled with units of concentration (in micrograms per litre) for plasma and units of amount (in micrograms) for dialysate and urine.

Covariate effects were not explored owing to the small sample size and because visual inspection of diagnostic plots showed no significant parameter–covariate relationships. However, the estimated clearances and volumes were standardized to a 70 kg adult human to allow for future comparison with other subpopulations, including extrapolation to infants and children. The VPC demonstrated good predictive performance, although a mismatch between the lower (5%) observed and predicted percentiles categorized two patients with a relatively high oseltamivir carboxylate clearance. A post hoc analysis of parameter estimates identified two subgroups of renal elimination: low urine clearance [shown by patients 1 and 3, with respective clearances of 0.195 and 0.382 l h⁻¹ (70 kg)⁻¹] and high urine clearance [patients 6 and 10, with respective clearances of 2.43 and 2.58 l h⁻¹ (70 kg)⁻¹]. These apparent differences accounted for the high BSV (117%; Table 2) in the clearance of oseltamivir carboxylate via renal elimination. However, inclusion of creatinine clearance as a covariate effect did not significantly improve the model and did not explain the random variability associated with clearance by this route (see Supplementary Figure S2). The VPC and goodness-of-fit plots indicated the ability of the model to predict adequately the central tendency and variability in the study cohort (see Figure 3 and Supplementary Figures S3–S5).

Figure 3.

Visual predictive check for oseltamivir (OP) in plasma (A), oseltamivir carboxylate (OC) in plasma (B), oseltamivir carboxylate in dialysate (C) and oseltamivir carboxylate in urine (D). The model showed good predictive performance and identified two patients with a relatively high clearance of oseltamivir carboxylate (indicated by the mismatch between the 5% observed and predicted percentiles in B). Corresponding plots in urine (D) differentiated these two (high urine-producing) patients from the other two urine-producing patients, who had lower estimated urinary clearance. , observed data; , predicated 50th percentile (shaded, 95% confidence interval); , observed 50th percentile; , predicated 5th and 95th percentiles (shaded, 95% confidence interval); , observed 5th and 95th percentile

Monte Carlo simulations

Initial dosing simulations at 75 mg oseltamivir and the present APD regimen showed that target exposures are achieved in anuric patients or those with low residual urine clearance, but not those with high renal elimination (Table 3). For this last group, target AUC_0–120 and C₁₂₀ exposures were attained with 75 mg at 0 and 48 h or with 30 mg daily over five consecutive days (see Supplementary Figure S6). The following three different APD regimens were also simulated: CAPD only (three sessions over 15 h during the day, then a single 9 h exchange overnight); intermediate (one CAPD exchange over 10 h before a 2 h break during the day, then three CCPD sessions over 9 h, then a 3 h break overnight); and the present more intensive regimen (see Supplementary Figure S1). For the 75 mg single dose, median AUC_0–120 and C₁₂₀ were approximately threefold above the defined target in anuric patients. Of the three regimens, simulated exposures were highest for CAPD only and lowest for the present intensive CAPD/CCPD combination, although only minor differences were observed. Table 3 provides a summary of the PTA for a range of oral oseltamivir dosing regimens in anuric patients and in patients with residual renal function prescribed different PD regimens.

Table 3.

Probability of target attainment of oseltamivir carboxylate exposures in plasma following a range of oseltamivir dosing and automated peritoneal dialysis regimens simulated for 1000 patients

Exposure variable	Urinary clearance	Dosing regimen	Probability of target attainment*
			Initial simulation	CAPD only†	Intermediate APD‡	Intensive APD§
Plasma AUC0–120	Anuric	75 mg single dose	1.000	1.000	1.000	1.000
		60 mg single dose	1.000	1.000	1.000	1.000
		45 mg single dose	1.000	1.000	1.000	1.000
		30 mg single dose	0.973	0.984	0.977	0.976
	Low urine¶	75 mg single dose	1.000	1.000	1.000	1.000
		60 mg single dose	1.000	1.000	1.000	1.000
		45 mg single dose	0.998	1.000	0.999	0.994
		30 mg single dose	0.773	0.818	0.784	0.759
	High urine**	75 mg single dose	0.194	0.199	0.195	0.190
		60 mg single dose	0.035	0.044	0.040	0.040
		45 mg single dose	0.000	0.002	0.002	0.001
		30 mg single dose	0.000	0.000	0.000	0.000
		75 mg (0 and 48 h)	0.957	–	–	–
		30 mg (five daily doses)	0.913	–	–	–
Plasma C120	Anuric	75 mg single dose	0.974	0.980	0.978	0.966
		60 mg single dose	0.957	0.970	0.960	0.948
		45 mg single dose	0.929	0.940	0.917	0.904
		30 mg single dose	0.773	0.812	0.714	0.727
	Low urine¶	75 mg single dose	0.674	0.700	0.668	0.649
		60 mg single dose	0.547	0.591	0.544	0.515
		45 mg single dose	0.348	0.408	0.383	0.307
		30 mg single dose	0.065	0.080	0.072	0.053
	High urine**	75 mg single dose	0.000	0.000	0.000	0.000
		60 mg single dose	0.000	0.000	0.000	0.000
		45 mg single dose	0.000	0.000	0.000	0.000
		30 mg single dose	0.000	0.000	0.000	0.000
		75 mg (0 and 48 h)	0.043	–	–	–
		30 mg (five daily doses)	0.905	–	–	–

Abbreviations are as follows: APD, automated peritoneal dialysis; AUC_0–120, area under the concentration–time curve from time 0 to 120 h postdose; BSV, between-subject variability; C₁₂₀, plasma concentration at 120 h postdose; CAPD, continuous ambulatory peritoneal dialysis; CCPD, continuous cycler-assisted peritoneal dialysis.

^*Proportion of simulated patients (n = 1000) achieving AUC_0–120 of 31 820 ng h ml⁻¹ or C₁₂₀ of 170 ng ml⁻¹.

^†Three CAPD sessions over 15 h during the day, followed by a single exchange of 9 h overnight.

^‡A single CAPD exchange over 10 h before a 2 h break during the day, and three CCPD sessions over 9 h followed by a 3 h break overnight.

^§Three CCPD sessions over 8 h during the day, and two CAPD exchanges of 16 h overnight.

^¶Simulated using a median urinary clearance of 0.289 l h⁻¹ (70 kg)⁻¹ (BSV 45%), calculated from the individual estimates [0.195 and 0.382 l h⁻¹ (70 kg)⁻¹] from patients 1 and 3.

^**Simulated using a median urinary clearance of 2.50 l h⁻¹ (70 kg)⁻¹ (BSV 10%), calculated from the individual estimates [2.43 and 2.58 l h⁻¹ (70 kg)⁻¹] from patients 6 and 10.

Safety

There were no deaths or premature withdrawals. At least one TEAE was reported for nine of the 10 patients in the safety population (90.0%). Most TEAEs were mild in intensity and were not due to oseltamivir. The most frequent TEAEs were gastrointestinal tract and nervous system disorders. Nausea was reported for one patient and vomiting for one patient.

There was at least one severe TEAE in three patients (30.0%; four events). A single case of hypoglycaemia occurred that was unrelated to the drug, while three other events (hypokalaemia, lethargy and pruritus) were initially mild but did not have their worst intensity recorded. These were marked as severe as per protocol. One TEAE (mild oesophageal reflux) was considered as possibly related to oseltamivir. There was one respiratory tract infection of moderate intensity that was serious, because the patient was hospitalized, but was considered unrelated to the study drug and resolved without sequelae.

No clinically significant haematology findings were observed. Biochemical abnormalities were reported for eight patients but none were considered of clinical significance. A clinically relevant low blood glucose level was reported in one patient due to diabetes and was unrelated to oseltamivir. There were considerable variations, but without any discernible trends over time, in mean and median haematology parameters, liver function tests, systolic and diastolic blood pressure, pulse and body temperature.

Discussion

Patients with ESRD are at increased risk of developing influenza-related complications [7,8], but limited information is available on the suitability of oseltamivir for treatment and prophylaxis in this patient group. Moreover, the availability of oseltamivir pharmacokinetic data generally is limited for patients who are also receiving APD prescription. The present study provides new information regarding oseltamivir dosing in patients receiving APD treatment. It provides nephrologists with clearer guidelines on the APD and drug dosing regimens required to achieve optimal therapy in this patient cohort. The analysis also confirms the observations from a prior study in CAPD patients [12].

In the present study, administration of a 75 mg dose of oseltamivir immediately before commencement of APD produced plasma concentrations at 120 h (301 ng ml⁻¹) that exceeded steady-state concentrations (160 ng ml⁻¹) in patients with normal renal function (Roche, data on file). These patients received the usually recommended oral dosing regimen of oseltamivir (75 mg twice daily). In addition, the plasma concentrations achieved at 168 h postdose (138 ng ml⁻¹) exceeded the recommended minimum concentration (C_min = 40 ng ml⁻¹) for a prophylaxis regimen (75 mg once daily) in patients with normal renal function (Roche, data on file). Both the clinical observations and the model simulations support that a 75 mg single oral dose produces sufficiently high AUC (PTA = 1.0) and C_min (PTA = 0.974).

The exposures of oseltamivir carboxylate in ESRD patients were considerably higher than those previously observed in subjects with normal renal function receiving the recommended treatment dose. In the present study population, the AUC_0–inf (93 800 ng h ml⁻¹) was fourfold greater than the AUC_0–last (21 752 ng h ml⁻¹) reported in patients with normal renal function administered oseltamivir 75 mg twice daily [12]. The higher exposures in the present patient cohort are a consequence of residual urinary elimination that is probably lower than would be expected with normal kidney function. The present exposures were in fact similar to an oseltamivir dose of 450 mg twice daily (Roche, data on file), which was well tolerated in subjects with normal renal function and is considered the upper limit of the safety margin [20]. Furthermore, adverse events such as nausea and vomiting, which are commonly associated with oseltamivir exposure [21], were undetected in ESRD patients.

Oseltamivir was well tolerated in the present study, and there were no deaths or premature withdrawals. Nausea was reported for one patient and vomiting for one patient. Most TEAEs were of mild intensity and were not considered to be due to the study medication; events that were possibly related to oseltamivir were resolved without sequelae. There were no consistent haematological or biochemical trends and no new safety signals. These findings are broadly consistent with those reported in a cross-sectional study of 333 patients with ESRD receiving oseltamivir [22]. The most common events were nausea, abdominal pain and dizziness. No gastrointestinal disturbances occurred, which are the most frequently reported adverse event following oseltamivir treatment [21,23].

We noted significant variability in renal clearance that was due to a mixture of subjects with low and high residual urine production. As noted earlier, five patients were anuric, while one urine producer was an outlier, with an output of up to 2776 ml day⁻¹. Model development allowed for the explicit identification of two populations of urine-producing patients, in addition to anuric subjects. The importance of renal clearance was illustrated by the model, which indicated that patients with high residual urine production would be unlikely to attain target exposures with the present oseltamivir dose and aggressive APD regimen. However, further investigation is required to improve the definition of optimized dosing strategies, owing to the limited (n = 10) patient sampling size in the present study population. Additional data are also necessary to identify relationships between individual parameters and covariate effects.

The present dialysis treatment included a significant cycler-assisted component that was designed to simulate an aggressive APD prescription. Nonetheless, administration of a 75 mg dose produced exposures at the upper limit of safety, thereby suggesting that less aggressive APD prescriptions could produce exposures that exceed tolerability. In contrast, the recommended 30 mg dose for CAPD [9,10] may produce lower plasma trough concentrations than achieved with 75 mg twice daily dosing in patients with normal renal function. These findings are implicated for efficacy and the emergence of resistance and support the need for exploration of intermediate oseltamivir doses, such as 45 or 60 mg. In the present analysis, modelling and simulation demonstrated that a 45 or 60 mg dose achieved moderate plasma concentrations of oseltamivir carboxylate at 120 h postdosing (C₁₂₀ = 180–230 ng ml⁻¹). These dosage regimens are potentially more relevant in patients receiving less aggressive APD modalities (e.g. CAPD only), for which greater exposures are expected than currently observed. Furthermore, this (lower) dose adjustment would reduce tolerability concerns, whilst simultaneously achieving target exposures that are acceptable (PTA > 0.8) in anuric patients or those with low residual renal function. Implementation into clinical practice is also achievable, because intermediate dose forms, such as 45 mg capsules, are already available.

The pharmacokinetic model developed was consistent with previous studies of first-order absorption and rapid first-pass metabolism to produce an active metabolite [16,18]. We also noted that the estimated rate constant for oseltamivir carboxylate absorption was slower than that for the parent prodrug, which is expected with the slow release of a polar metabolite from hepatocytes. The estimated clearance by CCPD was considerably higher (approximately twofold) than that via CAPD, consistent with a rapid cycling APD regimen [5]. While ESRD can produce changes in plasma protein concentrations, this is unlikely to have a significant impact on the pharmacokinetics, owing to the low plasma protein binding (3%) of oseltamivir carboxylate.

Our model was developed using an APD regimen that was likely to generate rapid oseltamivir clearance, producing inadequate antiviral exposures and efficacy, with the possibility of an increased risk of resistance. We developed a population pharmacokinetic model that can optimize dosing with APD treatment and we identified a subgroup of urine-producing patients with a significantly higher drug clearance than anuric subjects. This effect was not predicted by serum creatinine measurement or estimated creatinine clearance. While further investigation is required in patients with high residual renal function, this subpopulation may respond better to alternative dosing regimens (such as 75 mg oseltamivir prior to APD and a second dose after 48 h, or 30 mg daily for five consecutive days). However, the present 75 mg single oral dose appears suitable for APD patients with negligible or low residual renal function. These findings were recognized only with the application of modelling and simulation, which are strongly recommended in future studies.

CONCLUSIONS

In ESRD patients managed with APD, administration of a single 75 mg oseltamivir dose resulted in substantial hepatic first-pass metabolism to the active metabolite, oseltamivir carboxylate. The development of a population pharmacokinetic model resulted in a higher estimated clearance of metabolite by CCPD than via CAPD. The resulting exposures of oseltamivir carboxylate were within the established safety margin, albeit at the upper end. Despite this, oseltamivir was generally well tolerated, with no new safety signals. The analysis also identified a significant contribution of urinary clearance in patients with residual kidney function that could produce inadequate exposures, with a potential requirement for dose adjustment. However, robust dosing guidelines were not established in the present patient cohort owing to the limited sample size. A more rigorous modelling and simulation analysis is required to determine whether lower doses can provide sufficient exposures in this high-risk population and to evaluate fully the importance of residual renal function.

Competing Interests

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare the following for the previous 3 years: MG, MAK and PNM are employees of Roche; KP and CMK report grants from Roche outside of the submitted work; CRR is a former employee of Roche and is currently employed by d3 Medicine, which provides consulting services to Roche outside of the submitted work; RR reports grants received from Roche by Christchurch Clinical Studies Trust for the conduct of the clinical study.

Supporting Information

Additional Supporting Information may be found in the online version of this article at the publisher's web-site: