Elimination of Doripenem during Dialysis and Pharmacokinetic Evaluation of Posthemodialysis Dosing for Patients Undergoing Intermittent Renal Replacement Therapy

ABSTRACT

Doripenem is a broad-spectrum parenteral carbapenem with enhanced activity against Pseudomonas aeruginosa and Enterobacteriaceae. Current dosing regimens recommend the administration of 0.25 to 0.5 g once daily in patients undergoing intermittent renal replacement therapy. As patients are usually dialyzed thrice weekly, we aimed to investigate a 1-g posthemodialysis regimen, thus reducing treatment costs and enhancing patient compliance. A second objective of this trial was to describe the pharmacokinetics of intradialytic doripenem. Ten oliguric or anuric patients in need of intermittent renal replacement therapy were included in this trial. All patients suffered from a septic episode. The mean hemofilter clearance was 123.46 ± 42.03 ml/min, and the total body clearance between hemodialysis sessions was 16.79 ± 6.02 ml/min. The average prehemodialysis trough concentration was 2.4 ± 1.3 mg/liter, while the EUCAST resistance breakpoint for Enterobacteriaceae is set at 2 mg/liter. The interpatient variability was considerably higher than the intrapatient variability. Apart from one patient who suffered an allergic reaction, doripenem was tolerated well by all patients. Our data indicate that posthemodialysis administration of 1 g of doripenem results in sufficient plasma levels in anuric but not oliguric patients during the entire dosing interval. (This trial was registered with EudraCT under registration no. 2009-018010-18 and at ClinicalTrials.gov under registration no. NCT02018939.)

INTRODUCTION

Due to their activity against multidrug-resistant Gram-negative (MRGN) bacteria, carbapenems serve an important role in the treatment of hemodialysis patients, who are often multimorbid. In most cases, the patients receive multiple courses of different antimicrobial agents per year, which are not necessarily dosed sufficiently, as previous trials have shown, especially for patients with chronic kidney disease (CKD) (1–3). Doripenem is a parenteral, broad-spectrum carbapenem antimicrobial agent with increased in vitro activity against Pseudomonas aeruginosa versus that of other carbapenem antimicrobial agents (4). It is eliminated mainly via the kidneys (5).

In adult patients in need of intermittent renal replacement therapy (iRRT), the suggested dosage for doripenem ranges from 0.25 to 0.5 g once daily (6, 7). However, most patients undergoing iRRT are treated in an outpatient setting; therefore, a posthemodialysis antimicrobial regimen would be desirable for such patients, as well as preferable from an economic viewpoint (8). Hemodialysis dosing recommendations are extrapolated from standard dosing regimens for renally competent patients. General doripenem dosing recommendations, however, were increased from 500 mg every 8 h (q8h) to 1 g q8h after a review by the European Medicines Agency in 2012 (9). As a consequence, a higher-dose regimen than that previously recommended might also be appropriate for patients undergoing hemodialysis.

Posthemodialysis administration has been described for other beta-lactams, such as meropenem, cefazolin, cefepime, and cefpirome (10–15). We thus hypothesized that a 1-g posthemodialysis regimen for doripenem may provide a sufficient time that the free, unbound concentration of drug exceeds the MIC (fT_MIC) to allow for safe and efficient treatment of hemodialysis patients with doripenem. As doripenem shows a very low level of protein binding (approximately 8%), we assumed fT_MIC to be sufficiently similar to T_MIC to omit protein binding in the calculations (5). The aim of the present trial (EudraCT registration no. 2009-018010-18; ClinicalTrials.gov registration no. NCT02018939) was to determine and evaluate the “real world” pharmacokinetics of doripenem administered in a 1-g posthemodialysis regimen in septic patients in need of iRRT and to characterize the intradialytic elimination of the drug. To our knowledge, this is the first report describing the thrice-weekly posthemodialytic administration of a carbapenem.

RESULTS

Patients, iRRT setting, and tolerability.

Between August 2011 and June 2013, five female and five male patients were included in this trial. The average age was 61.8 years, and the average body weight was 77.7 kg. As a consequence, the average dose was 13.3 mg/kg of body weight. The demographic data for the cohort may be found in Table 1. Due to a type I allergic reaction shortly after administration of the drug, the treatment of patient 1 could not be continued, and the patient 1 values were excluded from the analysis. During hemodialysis session 1 (HD1), patient 3 suffered from very low blood pressure due to her septic condition. As a result, hemodialysis had to be paused several times, resulting in inaccurately low clearance rates. These values are given in the tabular overview but were otherwise excluded from analysis. For patient 6, only venous samples could be obtained during HD1, for technical reasons; as a result, only the total body clearance (CL_tot), not the prefilter/postfilter clearance, could be calculated. These events contributed to a smaller number of time points available for clearance calculations.

TABLE 1.

Demographic data for the patient cohort^a

Screening no.	Sex	Age (yr)	Body wt (kg)	BMI	Dose (mg/kg)	Membrane	Membrane area (m2)	kUF value
1	F	51	53	20.70	18.87	Baxter Xenium 210	2.1	80
2	M	65	75	22.40	13.33	Nikkiso FDY-210GW	2.1	63
3	F	83	87	31.02	11.56	Nikkiso FDX210GD	2.1	63
4	M	73	100	29.34	9.96	Nipro Sureflux 210UH	2.1	70
5	F	77	74	28.91	13.51	Fresenius FX80	1.8	53
6	M	20	67	23.15	14.95	Fresenius FX80	1.8	53
7	M	52	72	22.07	13.99	Gambro Polyflux 24S	2.4	77
8	F	60	77	29.88	13.07	Baxter Xenium+ H21	2.1	82
9	M	66	70	23.49	14.39	Nipro Sureflux 21 UX	2.1	46
10	F	71	104	35.99	9.62	Fresenius FX100	2.2	68
Mean ± SD		61.8 ± 17.0	77.7 ± 14.6	26.7 ± 4.8	13.3 ± 2.5		2.1 ± 0.2	65 ± 11.6

^aAbbreviations: BMI, body mass index; F, female; M, male; k_UF, ultrafiltration coefficient; SD, standard deviation.

Dialysis was performed as prescribed by the treating physicians. Membrane sizes ranged from 1.8 to 2.4 m², with an average size of 2.1 m². While mostly polymer (Polynephron, polyester polymer alloy [PEPA], polysulfone, or Polyamix) membranes were used, patients 4 and 9 received treatment with triacetate membranes. Ultrafiltration quotients ranged from 46 to 82, with an average value of 65.5. Typical blood flow rates used were 250 and 300 ml/min. The mean blood flow during HD1 was 280 ± 36 ml/min. The mean ultrafiltration rate was 709 ± 257 ml/h. The typical dialysate flow was 500 ml/min; only for patients 8 and 10 was a dialysate flow of 800 ml/min used. Five patients received a hemodialysis treatment. In five patients, hemodiafiltration with a mean substitution rate of 91 ± 20 ml/min was performed. Individual hemodialysis parameters during HD1 may be found in Table S1 in the supplemental material.

Noncompartmental pharmacokinetic analysis.

The mean doripenem peak value during HD2 to HD6 was 44.7 ± 19.5 mg/liter (n = 38 individual time points), while the mean prehemodialysis trough level was found to be 2.4 ± 1.3 mg/liter (n = 37). Patients 6 (75% of all prehemodialysis trough levels were <2 mg/liter), 9 (80% of levels were <2 mg/liter), and 10 (100% of levels were <2 mg/liter) showed mean trough levels of <2 mg/liter and contributed largely to the high interpatient variability of 49.8%. In contrast to this, the intrapatient variability was only 31.9%. Although none of these three patients exhibited atypically high interdialytic clearance, all three had comparably high residual urine volumes of only slightly less than 500 ml/24 h, whereas all other patients had 0 or close to 0 ml of residual urine volume. A logarithmic graph of the mean plasma concentrations during HD2 to HD6 was created (see Fig. 3). The volume of distribution (V) was 35.8 ± 10.8 liters (n = 9).

FIG 3.

Trial design and intra- and interdialytic pharmacokinetics of doripenem. The graph shows average prefilter concentrations of doripenem during the course of the trial. The positive ranges of standard deviations are indicated by the error indicators (whiskers). HD, hemodialysis session.

The average intradialytic total body clearance was 250 ± 43.2 ml/min (n = 9), while the average intradialytic prefilter/postfilter clearance calculated from plasma concentrations during HD1 was 123.5 ± 42.0 ml/min (n = 34 individual time points), with a drug half-life (t_1/2) of 88.98 ± 27.2 min (n = 9). All intradialytic values, as well as the area under the concentration-time curve from 0 h to infinity (AUC_0–inf), were calculated based on the doripenem plasma concentrations measured prefilter during HD1. Those concentrations, as well as the postfilter and dialysate/ultrafiltrate concentrations of doripenem during HD1, may be found in Fig. 1. In contrast to the intradialytic clearance, the interdialytic clearance was strongly reduced, with a total body clearance of 16.8 ± 6.0 ml/min (n = 29 individual hemodialysis sessions) and an average half-life of 13.2 ± 3.9 h (n = 29 individual hemodialysis sessions). Individual pharmacokinetic results for all patients are presented in Table 2.

FIG 1.

Intradialytic pharmacokinetics of doripenem. The graph shows average doripenem concentrations over time after iRRT commencement measured prefilter, postfilter, and at the dialysate port of the hemofilter during HD1. The positive ranges of standard deviations are indicated by the error indicators (whiskers). The doripenem administration intervals are marked with the administered doses.

TABLE 2.

Individual and average pharmacokinetic results^a

Screening no.	Prefilter AUC0–inf_intradialytic (mg · h/liter)	Prefilter AUC0–eod_intradialytic (mg · h/liter)	CLtot_intradialytic (ml/min)	CLtot_interdialytic (ml/min)	CLpre-postfilter (ml/min)	V (liters)	V (liters/kg)	t1/2_intradialytic (min)
2	51.47	44.24	323.82		141.0	35.0	0.47	68.2
3	57.34	34.97	260.68	13.69		35.9	0.42	69.0
4	74.08	44.71	225.00	18.78	157.0	62.4	0.62	152.3
5	94.35	62.63	176.65	17.03	124.7	20.2	0.27	77.9
6	59.80	48.82	278.72	16.25		33.6	0.50	81.6
7	61.69	54.46	270.16	16.82	128.6	30.5	0.43	74.8
8	67.08	55.41	248.47	13.76	119.6	37.1	0.49	104.5
9	82.19	60.74	202.79	13.77	140.2	39.2	0.56	110.4
10	71.23	65.93	233.99	24.12	135.9	28.1	0.27	62.1
Mean ± SD	68.8 ± 11.69	51.66 ± 9.31	250.03 ± 43.24	16.79 ± 6.02	123.46 ± 42.03	35.8 ± 10.8	0.44 ± 0.11	88.98 ± 27.76

^aAUC, area under the concentration-time curve; CL_tot, total body clearance; CL_{pre-postfilter}, clearance measured between the hemofilter inlet and outlet; V, volume of distribution; t_1/2, half-life. Mean values were calculated from values at all individual time points rather than from the average patient values shown in this table.

Population pharmacokinetic model.

A two-compartment model with linear elimination best described the intradialytic pharmacokinetics of doripenem, which is in line with the results from previous population pharmacokinetic analyses (16, 17). Intradialytic total clearance during HD1 was estimated to be 12 liters/h, with moderate interindividual variability (coefficient of variation [CV], 20%). Total volume of distribution accounted for approximately 30 liters and revealed higher variability between the patients (54.4% CV) (Table 3). A combined (additive and proportional) model best captured residual variability. The visual predictive check (VPC) indicated high predictive performance of the developed population model for the general trend and the variability between the patients (Fig. 2).

TABLE 3.

Parameter estimates for the intradialytic, intrapopulation pharmacokinetic model of doripenem^a

Parameter (unit)	Estimate (RSE [%])
Fixed-effects parameters
CL (liters/h)	12.1 (8.00)
V1 (liters)	19.9 (24.2)
Q (liters/h)	10.2 (34.4)
V2 (liters)	11.1 (13.6)
Interindividual variabilities
CL CV%	20.0 (37.9)
V1 CV%	54.4 (46.5)
Residual variabilities
Additive SD (mg/liter)	1.15 (36.3)
Proportional CV%	18.2 (14.9)

^aCL, clearance; V₁, central volume of distribution; Q, intercompartmental clearance; V₂, peripheral volume of distribution; CV, coefficient of variation; SD, standard deviation; RSE, relative standard error (RSE of interindividual and residual variability parameters are reported on an approximated standard deviation scale).

FIG 2.

Visual predictive check of the intradialytic population pharmacokinetic model of doripenem at HD1 (based on 1,000 simulations). Circles represent the observed data; lines represent the 10th, 50th, and 90th percentiles of the observed data; and shaded areas represent the 95th confidence intervals around the 10th, 50th, and 90th simulated percentiles.

DISCUSSION

Doripenem was eliminated to a large extent during iRRT. While there were differences between the individual patients, no large impact of different membrane sizes or materials was identified, as indicated by the similar intradialytic clearance rates (CL_{pre-postfilter}) for all patients. This may be attributed to doripenem already possessing good dialyzability with membranes with low ultrafiltration coefficients (k_UF), resulting in a comparably low impact of membranes with enhanced clearance capabilities. Clearance differences between hemodialysis and hemodiafiltration were negligible. Compared to an earlier report describing doripenem pharmacokinetics during hemodialysis, our patients were treated with higher blood flow rates (7). As a consequence of the higher dose administered, maximum doripenem plasma concentrations exceeded those presented earlier (67.0 ± 24.8 mg/liter versus 24.56 ± 11.20 mg/liter) (7). The prefilter/postfilter clearance values found in our trial are in line with those described by Tanoue et al. (123.5 ± 42.0 ml/min versus 117.78 ± 20.50 ml/min) (7). Compared to historic data for doripenem (83.83 ± 7.11 ml/min) or meropenem (79 to 81 ml/min), doripenem was cleared to a larger extent in our trial (15, 18). It should be noted, however, that mainly Cuprophan membranes were used in those trials. A more recent publication describes the intradialytic clearance of meropenem to be 5.78 ± 1.03 liters/h (equaling 96.3 ± 17.2 ml/min), which is still considerably below the rate of 123 ml/min found in our trial (19). In addition, the volume of distribution found in our patients exceeds that described for doripenem in the FDA approval reviews (16.9 ± 3.67 liters), as well as that of meropenem (18, 19). This finding most likely is owed to the septic state of the patients. In comparison to doripenem clearance values reached during continuous hemodiafiltration (36.5 ± 13.8 ml/min), doripenem clearance during intermittent hemodialysis is much higher (16). As a consequence, medical personnel should be advised that doripenem should be administered only after disconnection from the hemodialysis machine. Interestingly, the interdialytic clearance published in the FDA approval review (1.99 ± 0.906 liters/h) far exceeds that found in our subjects (1.01 ± 0.36 liters/h) (18).

Apart from patient 1, who suffered an allergic reaction, doripenem was well tolerated by all participants. Although the postinfusion plasma levels were high, no nausea or neurological complications were observed. No signs of drug accumulation were found; as such, we have no reservations against prolonged doripenem treatment if needed.

Prehemodialysis trough levels were comparably low if Enterobacteriaceae or Pseudomonas spp., which are classified as doripenem susceptible for MICs of up to 2 mg/liter (20), were targeted. For drug dosing, these trough levels should be considered the most important target values, as intradialytic clearance rates presented in this trial probably were underestimated due to the nature of their calculation. In contrast to severely ill patients on continuous hemodialysis, for non-critically ill patients, a pharmacokinetic goal of a targeted MIC at 35 to 40% of the dosing interval may be deemed sufficient (21, 22). A higher T_MIC in plasma might, however, be required to achieve sufficient drug concentrations at the site of infection (23–25). Doripenem plasma concentrations within this trial sufficiently exceeded the MIC in patients with a residual urine volume of <400 ml/min. In patients with an increased residual urine volume, higher doripenem doses or more frequent administration might be warranted. Thus, further investigation with a larger patient group and a systematic pharmacokinetic/pharmacodynamic target attainment analysis with a population pharmacokinetic model are warranted to confirm the findings.

Conclusions.

Our data support our initial hypothesis that thrice-weekly posthemodialysis administration of 1 g doripenem may result in sufficient plasma levels in anuric patients during the entire dosing interval. This is a feasible and, based on previous literature, cost-effective procedure. Intradialytic clearance with high-flow membranes is high enough to avoid drug accumulation. Furthermore, doripenem shows good tolerability in patients following this regimen.

MATERIALS AND METHODS

Patient eligibility.

Ten septic patients treated with intermittent hemodialysis three times a week (Monday, Wednesday, and Friday) due to end-stage renal disease were included in this open-label prospective trial. Patients provided oral and written informed consent before any trial actions were performed. All patients produced <500 ml urine per 24-h period. Membranes and flow rates were unchanged in relation to those for the standard treatment of the respective patient. The protocol was approved by the independent ethics committee of the Medical University of Vienna and by the relevant authorities.

Drug administration and sampling.

To assess the intradialytic pharmacokinetics of doripenem, each patient received an intravenous infusion of 1 g doripenem over 30 min before the initiation of hemodialysis session 1 (HD1) (see Fig. S1 in the supplemental material). Venous blood was drawn before and at the end of the infusion (−0.5 and 0 h). At 0 h, hemodialysis commenced. During HD1, prefilter (at the “arterial” inlet port of the hemofilter) and postfilter (at the “venous” outlet port of the hemofilter) blood samples as well as dialysate/ultrafiltrate (from the “dialysate” port of the hemofilter) samples were drawn at 0.5, 1, 2, 3.5, and 4 h. At the end of HD1 (4 h), another venous blood sample was drawn, and each patient received 500 mg doripenem intravenously (over 30 min) to compensate for possible intradialytic drug filtration (Fig. 1 and 3). After this, no additional blood draw was performed. In the following course of the trial, prefilter doripenem levels (equivalent to the patients' plasma concentrations of drug) were measured before and after HD2 to HD6 to assess the interdialytic clearance (Fig. 3). After HD2 to HD6 (each lasting 4 h), each patient received 1 g of doripenem intravenously over 30 min, followed by a venous blood draw (from a different intravenous cannula than the one that was used for doripenem infusion) 30 min after the end of the infusion for the assessment of doripenem peak levels. Interdialytic clearance was calculated as dose/AUC_0–nextHD for HD2 to HD6.

Samples were drawn into lithium-heparin-containing blood collection tubes and centrifuged at 3,000 × g for 10 min. The resulting plasma was transferred to cryotubes and stored at −80°C.

Drug assay.

The concentrations of doripenem in plasma and dialysate samples were determined by high-pressure liquid chromatography (HPLC). Frozen plasma samples from the patients were thawed at room temperature and then centrifuged at 13,000 × g for 5 min. Briefly, after the addition of 200 μl methanol to 100 μl plasma, the samples were centrifuged (13,000 × g for 5 min), and 80 μl clear supernatant was injected into the HPLC column. Microdialysate samples (10 μl) were injected into the column without any prior precipitation procedure. The determination of the doripenem concentration was performed using a Dionex UltiMate 3000 system (Dionex Corp., Sunnyvale, CA, USA) with UV detection at 298 nm. Chromatographic separation was carried out at 35°C on a Hypersil BDS-C₁₈ column (5 μm × 250 mm × 4.6-mm internal diameter [ID]; Thermo Fisher Scientific, Inc., Waltham, MA, USA) preceded by a Hypersil BDS-C₁₈ precolumn (5 μm × 10 mm × 4.6-mm ID). The mobile phase consisted of 0.1% acetic acid-methanol (90:10 [vol/vol]) at a flow rate of 1.0 ml/min. Calibration of the chromatogram was accomplished using the external standard method. Linear calibration curves were calculated from the peak areas of doripenem compared to those for the external standard by spiking drug-free human plasma and microdialysate samples with standard solutions of doripenem to obtain a concentration range of 0.01 to 10 μg/ml (average correlation coefficients of >0.99). The limit of quantification for doripenem in plasma and ultrafiltrate/dialysate was 0.01 μg/ml. Coefficients of accuracy and precision for this compound were <8.7%.

Noncompartmental pharmacokinetic analysis.

Methods for pharmacokinetic analysis using commercially available software (Kinetica 3.0; Innaphase Philadelphia, USA) have been described previously (26, 27). The area under the concentration-time curve (AUC) was determined by noncompartmental analysis using the linear trapezoidal rule. The elimination half-life was calculated as follows: t_1/2 = ln 2/k_el, where k_el (elimination rate constant) is the slope of the decreasing part of the concentration-time curve. The total body clearance (CL_tot) was determined by calculating the dose/AUC_0–inf value. The volume of distribution (V) was calculated as follows: V = CL_tot/k_el.

The filter clearance was calculated based on pre- and postfilter concentrations by using previously published formulae (16, 28), as follows: CL_pre/post = (C_pre − C_{post corr})/C_pre × BFR × (1 − Hct) and C_{post corr} = C_post × (BFR − FRR)/BFR, where BFR is the blood flow rate, C_pre is the doripenem plasma concentration measured at the hemofilter inlet, C_post is the doripenem concentration at the hemofilter outlet, FRR is the fluid removal rate during hemodialysis, and Hct is the hematocrit value of the blood.

Population pharmacokinetic model.

For characterization of the intradialytic doripenem pharmacokinetics in more detail, a population pharmacokinetic model was developed based on the doripenem prefilter concentrations extensively sampled during HD1. Population pharmacokinetic modeling was performed in NONMEM 7.3 (Icon Development Solutions, Ellicott City, MD, USA) and Perl Speaks NONMEM 4.6.0 (Uppsala University, Uppsala, Sweden), using first-order conditional estimation with interaction. Different compartment disposition models were investigated and parameterized by clearance (CL), volumes of distribution (V_x), and potentially intercompartmental clearance (Q) terms. Interindividual variability was implemented in a stepwise procedure by using exponential functions, hence assuming a log-normal distribution of the individual pharmacokinetic parameters. For incorporation of residual unexplained variability, different models (additive, proportional, and combined) were investigated. Model performance was assessed by the objective function value (OFV), parameter precision, goodness-of-fit plots, and visual predictive checks (VPC; n = 1,000 simulations).