Xia Li (1), Johannes Zander (2), Uwe Fuhr(1), Michael Zoller (2), Max Taubert (1)
(1) Cologne University Hospital, Department I of Pharmacology, Cologne, Germany; (2) Hospital of the Ludwig-Maximilians-University of Munich, Munich, Germany.
Objective:
For critically ill patients, physiological changes result in altered ciprofloxacin pharmacokinetics, possibly causing inadequate drug exposure. We aimed to identify those covariates which are the best predictors of ciprofloxacin pharmacokinetics concerning kidney and liver function in ICU patients.
Method:
A population pharmacokinetic model was built based on 444 ciprofloxacin plasma concentrations collected on four consecutive days from 15 critically ill patients with severe infections receiving standard dosing of ciprofloxacin in a prospective observational study. Relationships between ciprofloxacin clearance and renal, hepatic and biometric covariates were evaluated independently on study days 1 to 4 to assess whether a covariate was consistently related to clearance throughout the treatment course. For covariates which reduced the AIC, the stability of the relationship was confirmed by evaluating the 95% confidence intervals (CI) of the respective covariate parameters as obtained from a bootstrap with 1,000 samples [1]. A covariate was introduced into the model if the related 95% CIs did not include zero on at least three out of four treatment days. Subsequently, the relevance of covariates was assessed by evaluating the magnitude of change in clearance compared to the typical population clearance that was explained by the covariate. The identified covariates were then combined in the final covariate model.
Result:
A two-compartment model with linear elimination and a combined error model appropriately described the pharmacokinetics of ciprofloxacin. Clearance and central volume of distribution increased from the first (16.2 [13.4 – 19.7] L/h and 24.2 [13.9-36.4] L) to the fourth (20.9 [17.8 – 25.8] L/h and 33.5 [21.3-48.8] L) study day (median [95% CI] from bootstrap). The peripheral volume and the inter-compartmental clearance were 83.2 (74.1-94.8) L and 71.2 (46.2-93.9) L/h, respectively (median and 95% CI). Bilirubin (Eq. 1), age (Eq. 2), and sex (Eq. 3) were identified as covariates consistently related to clearance throughout the treatment course. In the final covariate model, the unexplained inter-individual variability of clearance was reduced from 55% (base model) to 22%.
Eq. 1 CLBILI=(BILI/1.85)**θeffect of BILI
Eq. 2 CLAGE=1+(49-AGE)*θeffect of AGE
Eq. 3 CLSEX=1+(1-SEX)* θeffect of SEX
Eq. 4 CL= θCL * CLBILI*CLAGE*CLSEX
Equation 1 to 4, covariate equations describing the identified relationships between bilirubin (BILI, mg/dL), age (AGE, years), sex (SEX, 1 = male) and ciprofloxacin clearance. The effect of total plasma bilirubin concentration on the clearance θ effect of BILI was -0.25(-0.374- -0.138), the effect of age on the clearance θ effect of AGE was 0.0156(0.0021-0.0208), and the effect of sex on the clearance θ effect of SEX was -0.413(-0.559- -0.214).
Conclusion:
According to our research, age, sex, and bilirubin might be more relevant to predict ciprofloxacin pharmacokinetics in critically ill patients, rather than creatinine clearance calculated by the Cockcroft-Gault equation[2] as described previously [3].
References:
[1] M. R. Gastonguay, “A Full Model Estimation Approach for Covariate Effects: Inference Based on Clinical Importance and Estimation Precision.,” AAPS J., vol. Abstract, no. W4354, 2004.
[2] D. W. Cockcroft and M. H. Gault, “Prediction of creatinine clearance from serum creatinine.,” Nephron, vol. 16, no. 1, pp. 31–41, 1976.
[3]D. Khachman et al., “Optimizing ciprofloxacin dosing in intensive care unit patients through the use of population pharmacokinetic-pharmacodynamic analysis and Monte Carlo simulations,” J. Antimicrob. Chemother., vol. 66, no. 8, pp. 1798–1809, Aug. 2011.
Reference: PAGE 27 (2018) Abstr 8594 [www.page-meeting.org/?abstract=8594]
Poster: Methodology - Covariate/Variability Models