Population pharmacokinetics of unbound cefazolin in infected hospitalized patients requiring intermittent high-flux haemodialysis: can a three-times-weekly post-dialysis dosing regimen provide optimal treatment?

Abstract

Objectives: To describe the population pharmacokinetics of cefazolin in infected hospitalized patients requiring intermittent haemodialysis (IHD).

Methods: This prospective population pharmacokinetic study was conducted in IHD patients prescribed cefazolin 2 g three times weekly. Plasma samples were collected at prespecified timepoints and assayed for total and unbound concentrations using validated LC. Pharmacokinetic modelling and dosing simulations were performed using Pmetrics®. PTA in plasma suitable for MSSA (unbound trough concentrations of ≥2 mg/L for the final 24 h of a 72 h interval) were simulated for different dosing regimens. A PTA of ≥95% was deemed acceptable.

Results: A total of 260 cefazolin concentrations (130 total, 130 unbound) were collected from 16 patients (14 female) with a median age of 51 years. The median (IQR) pre-dialysis unbound cefazolin concentration for a 3 day dose interval trough was 17.7 (13.5-31.4) mg/L. The median (IQR) unbound fraction was 0.38 (0.32-0.46). The lowest pre-dialysis unbound concentration was 9.1 mg/L. A two-compartment model with a complex protein-binding component adequately described the data. The mean unbound cefazolin CL during IHD was 16.4 ± 4.26 L/h, compared with 0.40 ± 0.19 L/h when dialysis was off. Duration of time on haemodialysis (TOH) was the only covariate supported in the final model. The 2 g three-times-weekly regimen was associated with a PTA of 99.7% on dosing simulations to maintain unbound concentrations of ≥2 mg/L with TOH of 6 months. The 1 g three-times-weekly post-dialysis was associated with a PTA of 95.4%.

Conclusions: A 2 g three-times-weekly post-dialysis cefazolin regimen is supported for MSSA infections.

Figure 1.

Observed total (grey cross) and unbound (black cross) cefazolin concentration–time data are presented with predicted lines of the final model (total, grey line; unbound, black line) for each individual patient.

Figure 2.

Observed-versus-predicted goodness-of-fit plots for unbound and total cefazolin concentration (open circles). (a) Population- and individual-predicted unbound cefazolin concentration; (b) population- and individual-predicted total cefazolin concentration. The cefazolin concentration shown are in mg/L. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

Figure 3.

Residual diagnostic plots for unbound cefazolin concentrations. From left to right: weighted residual error versus predicted concentration; weighted residual error versus time; weighted residual error frequency histogram. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.