In vivo / in vitro Correlation of Pharmacokinetics of Gentamicin, Vancomycin, Teicoplanin and Doripenem in a Bovine Blood Hemodialysis Model

Abstract

Background: Elimination of a drug during renal replacement therapy is not only dependent on flow rates, molecular size and protein binding, but is often influenced by difficult to predict drug membrane interactions. In vitro models allow for extensive profiling of drug clearance using a wide array of hemofilters and flow rates. We present a bovine blood based in vitro pharmacokinetic model for intermittent renal replacement therapy. Methods: Four different drugs were analyzed: gentamicin, doripenem, vancomicin and teicoplanin. The investigated drug was added to a bovine blood reservoir connected to a hemodialysis circuit. In total seven hemofilter models were analyzed using commonly employed flow rates. Pre-filter, post-filter and dialysate samples were drawn, plasmaseparated and analyzed using turbidimetric assays or HPLC. Protein binding of doripenem and vancomycin was measured in bovine plasma and compared to previously published values for human plasma. Results: Clearance values were heavily impacted by choice of membrane material and surface as well as by dialysis parameters such as blood flow rate. Gentamicin clearance ranged from a minimum of 90.12 ml/min in a Baxter CAHP-170 diacetate hemofilter up to a maximum of 187.90 ml/min in a Fresenius medical company Fx80 polysulfone model (blood flow rate 400 ml/min, dialysate flow rate 800 ml/min). Clearance of Gentamicin vs Vancomicin over the F80s hemofilter model using the same flow rates was 137.62 mL vs 103.25 ml/min. Doripenem clearance with the Fx80 was 141.25 ml/min. Conclusion: Clearance values corresponded very well to previously published data from clinical pharmacokinetic trials. In conjunction with in silico pharmacometric models. This model will allow precise dosing recommendations without the need of large scale clinical trials.

Keywords: antimicrobial agent; bovine blood; chronic hemodialysis; clearance; pharmacokinetics.

FIGURE 1

Doripenem drug concentrations during in vivo and in vitro hemodialysis. In vitro timepoints were multiplied by a factor of 11.36 which is the ratio between Vd in vivo and in vitro. Whiskers indicate standard deviation. In vivo data has been published in 2018 [16].

FIGURE 2

Best-fit straight-line correlation graph between previously published in vivo hemodialysis clearance rates and the corresponding results from the in vitro experiments. Number of Points 6, Degrees of Freedom 4, Residual Sum of Squares 2,43,25,364, Pearson’s r, 0,98667, Adj. R-Square 0,96689

FIGURE 3

Different Gentamicin clearances found in different membranes and flow rates. Membranes: CAHP 170: Baxter high performance cellulose diacetate membrane, 1.7 m² surace area; SF17UX: Nipro Surflux-17UX high performance cellulose triacetate membrane, 1.7 m² surface area; SF21UX: Nipro sureflux-21UX high performance cellulose triacetate membrane, 2.1 m² surface area; SF21E: Nipro Sureflux-21E balanced medium-flux cellulose triaceate membrane, 2.1 m² surface area; F80s: Fresenius medical care F80s, high flux polysulphone membrane, 1.8 m² surface area; FX80: Fresenius medical care FX 80classix high flux polysulphone membrane, 1.8 m² surface area. Flow rates are given below the membrane identifier: first number represents bloodflow in ml/min, second number dialysate flow in mL/min. Standard deviation is identified by whiskers.