Electronic Supplementary Material
Clinical Pharmacokinetics
A PBPK perspective on the clinical utility of albumin-based dose adjustments in critically ill patients
T’jollyn H.1, Vermeulen A.1,2, Van Bocxlaer J.1, Colin P.1
1 Laboratory of Medical Biochemistry and Clinical Analysis, Faculty of Pharmaceutical Sciences, Ottergemsesteenweg 460, 9000 Ghent, Belgium
2 Quantitative Sciences, Janssen Research & Development, a Division of Janssen Pharmaceutica NV, Turnhoutseweg 30, 2340 Beerse, Belgium
Contact: ;
- Drug binding behavior to albumin
In the current analysis, the KaPR(the association constant between drug-protein complex) was provided as the reciprocal value of the dissociation constant (KdALB), which is provided by the user. In turn, the KaPR is used to calculate drug binding in plasma (fup, EqA1) and tissues (Kpu, EqA2), depending on the molar concentrations of albumin on either side of the vascular wall ([ALB]plasma for the albumin concentration in plasma and [ALB]tissue for the albumin concentrations in the tissues). In addition, tissue binding is different across tissues since fractions of intracellular (fIW) and extracellular water (fEW), neutral lipids (fNL), phospholipids (fNP) and the albumin concentrations [ALB]T are tissue-specific {Rodgers, 2006 #15}. Instead of providing absolute albumin concentrations for every tissue ([ALB]tissue in EqA2), the tissue concentrations of albumin are governed by the ‘albumin partition coefficient’ (KpALB). This tissue-specific partition coefficient is determined by the ratio of the (steady-state) concentrations of albumin in the tissues and in plasma (Eq A3). Multiplication of KpALB with the albumin concentration in plasma, provides the albumin concentration in the tissues.
Eq A1
With fup as unbound drug fraction in plasma, KaPR as the association constant between drug-protein complex, [ALB]plasma as the concentration of albumin in the plasma
Eq A2
With Kpu as the ratio of total tissue concentration over unbound plasma concentration, X and Y represent the drug’s concentration difference due to pH differences between intracellular and extracellular milieu, fIWas fraction of intracellular water, fEW as fraction extracellular water, fNL as fraction neutral lipids, fNP as fraction phospholipids, [ALB]tissue as the albumin concentrations in a specific tissue
Eq A3
- Construction of hypoalbuminemic population files in the Simcyp Simulator
In all 3 hypo-albuminemic (“HYPO”) scenarios investigated, the intravascular albumin concentration was assumed to drop by 50%. HYPO1 represents the ‘capillary leakage’, with 50% of the intravascular albumin mass redistributed into the extravascular tissue water, according to the different perfusion rates of the tissues (see Appendix table A-1).
Table A1: based on the cardiac output to the different tissues, the redistribution of albumin was calculated for the hypo-albuminemic case 1: “albumin leakage”.
HYPO2 represents the case where a 50% drop in intravascular albumin concentration is achieved without change in the extravascular concentrations (intravascular loss). Hence, the tissue to plasma albumin partition coefficient (KpALB) for every tissue is simply doubled, compared to the normal situation, since only the plasma albumin concentrations were halved. HYPO3 represents a situation in which 50% of the total albumin mass intra- and extravascularly is removed. As a consequence, in this scenario, the KpALB for every tissue remains unchanged compared to the normal situation.
In Simcyp, each hypo-albuminemic scenario was represented by a different population file. First, the concentration of plasma albumin in the ‘tissue composition’ tab was adjusted to half of the normal plasma albumin concentration. Second, the albumin partition coefficients (KpALB) were changed in the software in order to represent the different albumin distribution scenarios: NORMO (control case), HYPO1, HYPO2, and HYPO3. An overview of these user-provided KpALB’s is given in Table A2. In order to focus specifically on the influence of plasma protein binding on unbound and total PK, we did not assume any fluid extravasation from intravascular to extravascular (third spacing).
In addition, for the simulation study a monoprotic acid with pKa 12 was chosen to represent a neutral compound. The reason for selecting a neutral compound is that with charged molecules, drug distribution is highly dependent on the net charge the molecule carries. Since the primary focus of this simulation exercise was on drug elimination, a neutral compound was selected for the sake of clarity. The reason for selecting in the PBPK software “a monoprotic acid with pKa 12” rather than the “neutral” compoundin the PBPK software is more technical. In the PBPK software used, neutral compounds are assumed to bind primarily to lipoproteins. Acids and bases will bind to albumin depending on the user-provided KD. However, since we wanted a neutral compound for the simulations, an acidic compound with pKa 12 was created so that no ionization occurs in the physiological pH range. This way, we avoid tissue retention/repulsion due to charge effects and ion-trapping.
1