Pharmacokinetics of epinephrine in patients with septic shock: modelization and interaction with endogenous neurohormonal status

Abstract

Introduction: In septic patients, an unpredictable response to epinephrine may be due to pharmacodynamic factors or to non-linear pharmacokinetics. The purpose of this study was to investigate the pharmacokinetics of epinephrine and its determinants in patients with septic shock.

Methods: Thirty-eight consecutive adult patients with septic shock were prospectively recruited immediately before epinephrine infusion. A baseline blood sample (C0) was taken to assess endogenous epinephrine, norepinephrine, renin, aldosterone, and plasma cortisol levels before epinephrine infusion. At a fixed cumulative epinephrine dose adjusted to body weight and under steady-state infusion, a second blood sample (C1) was taken to assess epinephrine and norepinephrine concentrations. Data were analyzed using the nonlinear mixed effect modeling software program NONMEM.

Results: Plasma epinephrine concentrations ranged from 4.4 to 540 nmol/L at steady-state infusion (range 0.1 to 7 mg/hr; 0.026 to 1.67 microg/kg/min). A one-compartment model adequately described the data. Only body weight (BW) and New Simplified Acute Physiologic Score (SAPSII) at intensive care unit admission significantly influenced epinephrine clearance: CL (L/hr) = 127 x (BW/70)0.60 x (SAPS II/50)-0.67. The corresponding half-life was 3.5 minutes. Endogenous norepinephrine plasma concentration significantly decreased during epinephrine infusion (median (range) 8.8 (1 - 56.7) at C0 vs. 4.5 (0.3 - 38.9) nmol/L at C1, P < 0.001).

Conclusions: Epinephrine pharmacokinetics is linear in septic shock patients, without any saturation at high doses. Basal neurohormonal status does not influence epinephrine pharmacokinetics. Exogenous epinephrine may alter the endogenous norepinephrine metabolism in septic patients.

Figure 1

Goodness-of-fit plot for the final model, observed vs. model-predicted epinephrine plateau concentrations. The prediction of the epinephrine plateau concentration at steady state infusion rate is: C_plateau (nmol/L) = (rate of infusion + R0)/(127 × (BW/70)^0.60 × (SAPS II/50)^-0.67) where R0 (nmol/hr) is the baseline rate of epinephrine infusion rate, BW (kg) is the body weight, and SAPS II is the severity score (new simplified acute physiology score) at intensive care admission.

Figure 2

Goodness-of-fit plot for the final model, normalized prediction distribution errors. The upper frame shows normalized prediction distribution errors (npde) vs. duration of epinephrine perfusion (delay C₀ to C₁) and the lower frame npde vs. model-predicted concentrations. The npde distribution was not significantly different from normality (P = 0.10 by Shapiro-Wilks test). Npde statistics are based on estimates of unbiased means and variances of the observations using 500 Monte Carlo simulations of the final model (the calculations include a de-correlation step of the prediction errors).