Pharmacokinetics During Hemodialysis and Hemodiafiltration: Dialysis Clearance is more important than Dialysis Efficiency - Application to Amikacin

Laila Nassar(1), Daniel Kurnik(1,2), Nick Holford(3)

(1) Rappaport Faculty of Medicine, Technion – Israel Institute of Technology, Haifa, Israel (2) Section of Clinical Pharmacology and Toxicology, Rambam Health Care Campus, Haifa, Israel (3) Department of Pharmacology and Clinical Pharmacology, University of Auckland, New Zealand

Objectives: 

Intermitted dialysis is the most common renal replacement therapy for patients with end stage renal disease. During hemodialysis (HD) the toxins in the blood are eliminated by diffusion through the dialyser membranes [1]. Diffusion of a solute depends on its molecular weight (mw), and larger molecules (>500 Da) are not effectively cleared by diffusion [2]. Hemodiafiltration (HDF) combines both diffusion and convection and eliminates medications more efficiently than HD when mw is over 500 Da [3].

Amikacin is an aminoglycoside with mw of 586 Da, thus expected to be eliminated by both convection and diffusion. Yet, there is little information on the effect of HDF on drug elimination and no data for amikacin. Current dosing regimens are largely based on clearance (CL) parameters estimated from HD [4], which can be expected to underestimate the additional elimination by HDF, thus increasing the risk of sub-therapeutic exposure [3]. This raises the question whether previous dosing recommendations are still relevant today [5]. To answer this question, theories of dialysis performance [6] and dialysis efficiency (DE) need further exploration.

In the last 20 years, patients with normal renal function have been receiving amikacin once daily instead of multiple doses a day, which achieves higher peak concentrations [7-10] while maintaining the same AUC as multiple daily dosing [11]. This regimen is as effective in treating infections and is associated with a lower rate of adverse effects compared to multiples doses a day [12-15].

For patients on intermittent dialysis, amikacin is administered as a single dose given after each dialysis session [4]. This dosing regimen is expected to yield lower peak and higher trough concentrations than patients with normal renal function receiving a higher once daily dose. HDF is expected to result in a higher dialyser CL.

The objectives of this work are to:

1. Devise a framework for describing DE and its connection to organ extraction ratio (ER)
2. Develop a population pharmacokinetic (PK) model of amikacin during HD or HDF
3. Quantitate dialyser CL and DE of amikacin during HD or HDF
4. Propose alternative dosing regimens of amikacin in patients receiving renal replacement therapy using HD or HDF

Methods: 

The theory describing dialyser blood CL (based on concentration at the dialyser input, (CL_B,i) and DE is largely based on the work of Michaels [6]. DE is calculated from the ratio of CL_B,i to dialyser blood flow (Q_B). DE can be shown to be equivalent to organ ER [16].

A single-center, prospective cohort study in adult patients treated with amikacin undergoing HD or HDF was performed. Blood samples were obtained before, during, and after dialysis. Patient characteristics, laboratory results, amikacin dosing regimen, amikacin concentrations and dialysis session characteristics were recorded. These records were used to develop a population PK model. PK parameters such as plasma CL (CL_P,i) were estimated based on the concentration at the site of elimination. Converting plasma (CL_P,i) to blood (CL_B,i) was done using the patients’ hematocrit (HCT), and population estimates of the partition ratio between blood and plasma (K_P) and the fraction unbound in plasma of amikacin f_up [17].

*f_up =1, K_P =0 for amikacin [11, 18]

Four models were explored to describe CL_B,i, i.e., fully empirical, fully mechanistic and two semi-empirical semi-mechanistic models.

Target exposure metrics were area under the plasma concentration time curve for 24 hours from the start of dialysis (AUC_24h) 285 mg*h/L, acceptable range 210-360 mg*h/L, peak plasma concentration of 52.5 mg/L (45-60 mg/L) and trough concentration 6 mg/L (4-8 mg/L). The exposure metrics and acceptable ranges were used to evaluate alternative dialysis dosing designs [19-21].

Data was analyzed using NONMEM [22] and Wings for NONMEM [23].

 Results: 

Twenty patients were studied based on 30 dialysis sessions (17 HDF and 13 HD), with a total of 188 plasma samples. A two-compartment model, with zero-order input and first-order distribution between the peripheral and the central compartment best fit the data. Cartridge type and coagulation in the dialysis cartridge at the end of dialysis were used as covariates. The objective function of the fully empirical (477.2) and the fully mechanistic (480.3) models were lower than the semi-empirical semi-mechanistic models (483 and 483.4). DE was 24% higher using HDF than HD based on the fully mechanistic model, however, dialyser blood CL, CL_B,i of amikacin using HDF was 62% higher than during HD. The higher performance comparing HDF with HD is because of the different blood flows associated with dialysis modality. Q_B for HDF was 21 L/h and 18 L/h for HD, making CL_B,i performance (DE*Q_B (Equation 1)), larger for HDF compared to HD.

A fully mechanistic based model provides more certainty for extrapolation than empirical models and therefore it was used to simulate dosing regimens. The current recommended dosing regimen (5 mg/kg given after 4 hours of dialysis) achieves neither the target AUC nor the target peak with either HD or HDF. Administering 15 mg/kg amikacin (the typical dose with normal renal function) at the beginning of dialysis achieves higher AUC and peak than that achieved with the recommended regimen, for both HD and HDF, but still below the acceptable range. For HD, increasing the dose to 17.5 mg/kg achieves AUC, peak and trough within the acceptable range. For HDF, increasing the dose to 25 mg/kg achieves AUC and trough but not the peak within the acceptable range. Increasing the dose to 30 mg/kg achieves a peak within the acceptable range but leads to a higher trough (Table 1), which may be associated with toxicity [14].

Table 1: Simulated alternative dosing regimens. *Italic values are outside of the acceptable range, Bolded values are within the acceptable range

Conclusions: 

A theoretical analysis of dialyser clearance and dialysis efficiency has shown that dialyser clearance is more informative for quantitative comparison of HD and HDF elimination performance than dialysis efficiency. Increasing the dose at the beginning of the dialysis session to 17.5 mg/kg for HD would reach acceptable target exposure metrics. For HDF, a higher dose of 25 mg/kg at the beginning of dialysis is required.

The principles used to develop target exposure based dosing in patients on dialysis are not limited to amikacin but are generally applicable to other drugs. The consequences for dosing when using HDF compared with HD are better predicted by considering dialyser clearance rather than dialysis efficiency because of the different contribution of dialyser blood flow to elimination.

References:
[1] Basile, C., A. Davenport, and P.J. Blankestijn, Why choose high volume online post-dilution hemodiafiltration? Journal of Nephrology, 2017. 30(2): p. 181-186.
[2] Tattersall, J.E. and R.A. Ward, Online haemodiafiltration: Definition, dose quantification and safety revisited. Nephrology Dialysis Transplantation, 2013. 28(3): p. 542-550.
[3] Zandvliet, A.S., D.J. Touw, and E.L. Penne, Medication and Hemodiafiltration, in Hemodiafiltration: Theory, Technology and Clinical Practice. 2016.
[4] Allon, M., Treatment guidelines for dialysis catheter-related bacteremia: an update. American journal of kidney diseases : the official journal of the National Kidney Foundation, 2009. 54(1): p. 13-7.
[5] Roberts, D.M., et al., Clinical Pharmacokinetics in Kidney Disease: Application to Rational Design of Dosing Regimens. Clinical journal of the American Society of Nephrology : CJASN, 2018. 13(8): p. 1254-1263.
[6] Michaels, A.S., Operating parameters and performance criteria for hemodialyzers and other membrane-separation devices. Transactions - American Society for Artificial Internal Organs, 1966. 12: p. 387-92.
[7] Barclay, M.L. and E.J. Begg, Aminoglycoside Adaptive Resistance Importance for Effective Dosage Regimens. Drugs, 2001. 61: p. 713-721.
[8] Craig, W.A., Pharmacokinetic/pharmacodynamic parameters: rationale for antibacterial dosing of mice and men. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America, 1998. 26(1): p. 1-2.
[9] Daikos, G.L., et al., Adaptive resistance to aminoglycoside antibiotics from first-exposure down-regulation.The Journal of infectious diseases, 1990. 162(2): p. 414-20.
[10] Freeman, C.D., et al., Once-daily dosing of aminoglycosides: review and recommendations for clinical practice. The Journal of antimicrobial chemotherapy, 1997. 39(6): p. 677-86.
[11] Winter, M.E., Basic Clinical Pharmacokinetics. 5th Editio ed. 2010.
[12] Begg, E.J., M.L. Barclay, and S.B. Duffull, A suggested approach to once-daily aminoglycoside dosing.British journal of clinical pharmacology, 1995. 39(6): p. 605-9.
[13] Parker, S.E. and P.G. Davey, Once-daily aminoglycoside dosing. Lancet (London, England), 1993. 341(8841): p. 346-7.
[14] Rybak, M.J., et al., Prospective evaluation of the effect of an aminoglycoside dosing regimen on rates of observed nephrotoxicity and ototoxicity. Antimicrobial agents and chemotherapy, 1999. 43(7): p. 1549-55.
[15] ter Braak, E.W., et al., Once-daily dosing regimen for aminoglycoside plus beta-lactam combination therapy of serious bacterial infections: comparative trial with netilmicin plus ceftriaxone. The American journal of medicine, 1990. 89(1): p. 58-66.
[16] Nassar, L., Pharmacokinetic Model Development for Target Concentration Intervention, PhD thesis. 2024, Technion - Israel Institute of Technology: Not yet published.
[17] Tozer, T.N., Concepts basic to pharmacokinetics. Pharmacology and Therapeutics, 1981. 12(1): p. 109-131.
[18] Lee, M.G., et al., Pharmacokinetics of drugs in blood I. Unusual distribution of gentamicin.Biopharmaceutics & Drug Disposition, 1981. 2(1): p. 89-97.
[19] Dang, L. and S. Duffull, Development of a Semimechanistic Model to Describe the Pharmacokinetics of Gentamicin in Patients Receiving Hemodialysis. The Journal of Clinical Pharmacology, 2006. 46(6): p. 662-673.
[20] Dager, W.E. and J.H. King, Aminoglycosides in intermittent hemodialysis: pharmacokinetics with individual dosing. Ann Pharmacother, 2006. 40(1): p. 9-14.
[21] Halouzkova, B.A., et al., Dosing of Aminoglycosides in Chronic Kidney Disease and End-Stage Renal Disease Patients Treated with Intermittent Hemodialysis. Kidney Blood Press Res, 2022. 47(7): p. 448-458.
[22] Beal, S., et al. NONMEM User’s Guides. (1989–2009). 2009.
[23] Holford, N.H.G., Wings for NONMEM Version 740 for NONMEM 7.3 and NONMEM 7.2. 2015.

Reference: PAGE 32 (2024) Abstr 10914 [www.page-meeting.org/?abstract=10914]

Oral: Clinical Applications