Background
Patients in the intensive care unit (ICU) are prone to develop bacterial infections, followed by an indication for antibiotic therapy. Due to altered pathophysiology in critically ill patients, finding the appropriate antimicrobial dosing is challenging with the risk of a drug over- or under-dosing, which may result in poor clinical outcome [
1]. Tigecycline (TGC) is a glycylcycline antimicrobial agent, approved in 2005 by the United States Food and Drug Administration (FDA) for the treatment of complicated intra-abdominal and skin-structure infections [
2‐
4]. It shows an expanded broad-spectrum activity against important and relevant sensible and multiresistant Gram-positive and Gram-negative bacteria such as (methicillin-resistant)
Staphylococcus aureus (MRSA), (vancomycin-resistant)
Enterococci (VRE) and (extended-spectrum β-lactamase-producing)
Enterobacteriaceae (ESBL) [
5]. TGC is considered as a last resort option for difficult-to-treat infections. However, data regarding suboptimal TGC dosing indicate a correlation with an increased risk of death [
6]. Consequently, several clinical studies are heading towards a high-dose regimen of TGC therapy as an approach to increase its efficacy [
6,
7].
On the other hand, understanding of the altered renal or hepatic function can be an alternative approach in adjusting antibiotic dosing. For instance, hepatic function plays a role in the clearance of TGC, as almost 60% of TGC is eliminated primarily via biliary excretion and approximately 20% is metabolized by the liver [
6,
8,
9]. Due to a lack of reliable liver function tests, it is difficult to obtain sufficient data guiding clinicians in TGC dose adjustment in critically ill patients with liver dysfunction.
The maximal liver function capacity (LiMAx) test has been recently introduced as a non-invasive diagnostic tool for determining acute liver failure in the intensive care medicine [
10]. It determines the enzyme activity of the liver based on a non-invasive breath test. Concerning antibiotic dosing of non-renal eliminated drugs, one recent study demonstrated a correlation between LiMAx and the pharmacokinetics (PK) of linezolid [
11].
In our study, we aimed to examine the impact of liver dysfunction on the PK of TGC in critically ill patients using the novel LiMAx test.
Discussion
The present study investigated the effect of liver failure on TGC PK for the first time. Based on recently conducted trials, we established a novel strategy using the LiMAx test for quantifying hepatic dysfunction [
10,
15,
16]. Our findings showed a significant increase in TGC PK curves in patients with strong hepatic dysfunction (group A), compared to patients with normal liver function (group C). The results of the multivariate analysis confirmed this effect. Regarding liver function, only LiMAx and total bilirubin revealed a significant impact on TGC
Cmax.
TGC
Cmax in our patient groups ranged from 0.441 to 1.774 mg/L. They were within the range to those determined in previous clinical trials with healthy volunteers [
3,
17]. However, extrapolating data of healthy subjects into critically ill patients is a challenge due to altered pathophysiology in this specific patient population. Particularly, the role of liver dysfunction is not sufficiently described, so far [
1]. Moreover, limited data are published, illuminating the pharmacokinetic of TGC in critically ill patients.
The LiMAx test has previously shown promising results in quantifying hepatic function in sepsis. The investigators concluded that LiMAx was a reliable diagnostic tool for identifying liver failure in critically ill patients [
10]. In another study, LiMAx explained a reasonable part of the linezolid PK variability of critically ill patients regarding the degree of liver failure. The authors demonstrated a strong association between LiMAx and non-renal clearance of linezolid by reducing the interindividual variability from 46.6 to 33.6%. LiMAx was superior to other markers of organ failure such as creatinine clearance (CLCr), thrombocyte count, total bilirubin and GGT even though an opportunistic probe sampling with linezolid trough levels was chosen in this work [
11]. The results of the present study with a more sophisticated strategy reflecting the PK of TGC demonstrate that LiMAx may provide an adequate diagnostic tool for predicting high TGC plasma levels in patients with hepatic dysfunction. Particularly, the LiMAx results < 100 µg/kg/h should lead to increased attention from the physicians. In vitro, data of TGC revealed a protein binding of about 50–70% [
18]. A low protein-binding of TGC yields a high volume distribution in the different body compartments and may lead to an imprecise interpretation of TGC plasma levels. Hence, a dosage algorithm based on the therapeutic drug monitoring of the TGC plasma level may be challenging to establish. Studies investigating TGC levels in other human body fluids such as bile or ascites are further required.
LiMax also provided some superior insights into the dosage regimes used across the norm. Since December 2018, EUCAST recommends a 200 mg loading dose of TGC followed by 100 mg steady-state dosage in treating critically ill patients infected with pathogens resistant to all other classes of antimicrobials [
19]. In our clinical study, a standard dosage, 100 mg loading dose of TGC followed by 50 mg steady-state dosage has been used. We observed, in Gram-positive bacteria, that critically ill patients in group A (68.383) had a significantly higher AUC above the MIC values than patients in group C (25.827). Similarly, in Gram-negative bacteria, group A (16.412) had a higher AUC above the MIC values than patients in group C (7.748). In such circumstances, EUCAST recommended dosage could lead to an increase in the risk of developing resistance towards TGC.
Another dynamic liver function test is the ICG-PDR, which is more widely used in clinical settings than the LiMAx test. Several authors investigated ICG-PDR in different conditions. The majority of authors came to the conclusion that ICG-PDR may not accurately measure liver dysfunction in sepsis due to complex ICG kinetics in liver disease and temporary redistribution into extrahepatic-extravascular tissues [
20]. Moreover, ICG elimination seems to be severely influenced by the splanchnic perfusion and is inhibited by hyperbilirubinemia, other anionic substances and acute cholestasis without evidence of changes in hemodynamic or morphology of hepatocytes [
21‐
24]. In addition, a recent study investigating liver function in sepsis showed a superiority of the LiMAx against the ICG-PDR in terms of quantifying liver dysfunction in critically ill patients [
10]. Such findings lead to the conclusion not to use the ICG-PDR in the recent study.
Static liver function parameters, such as AST, ALT, and GGT, failed to predict the degree of liver dysfunction accurately [
25]. The results in the present study confirmed these findings. Since the values of lactate, INR, platelet count and ALP differed significantly between groups A and C, these parameters failed to predict the variability of TGC
Cmax in the multivariate analysis. These findings are consistent with several studies, which identified only total bilirubin as a significant covariate to describe liver function [
26‐
28]. Other authors concluded that INR is a reliable diagnostic tool to define liver failure [
29]. INR can be influenced by disseminated intravascular coagulation or secondary hemorrhages, which are common complications in critically ill patients. A previous study in patients after major abdominal surgery showed similar postoperative progress of INR readouts compared with LiMAx values, while total bilirubin failed in predicting liver failure [
15]. These heterogeneous results of different studies investigating the accuracy of static liver function parameters in defining liver function, point out weak reliability of these parameters in this specific issue.
Besides the described dynamic and static liver function tests, other tools are introduced to define liver failure. Korth-Bradley et al. described a significant increase in TGC
Cmax and AUC in patients with advanced liver cirrhosis. The diagnostic tool used to define the degree of liver cirrhosis was the Child–Pugh score [
30]. However, the Child–Pugh score in critically ill patients may not be suitable to describe liver dysfunction reliably [
31]. In a systematic review, Cholongitas et al. mentioned that the Child–Pugh score in ICU patients can respond rapidly under treatment, resulting in an increase from class C to A. This apparent accelerated improvement might result in adjusting antibiotic dosage unnecessarily with the possible consequence of overdosing and side effects [
32]. Hence, guiding antibiotic therapy according to the Child–Pugh score in critically ill patients appears to be imprecise [
33]. Based on these data, we decided not to use the Child–Pugh score in the present study.
Another tool targeting liver failure is the MELD score. Initially, the MELD score was evaluated for patients undergoing a transjugular intrahepatic portosystemic shunt [
34]. The current version includes three objective variables (total bilirubin, INR and creatinine) and is simple to assess. The MELD score is predominantly used to prioritize the receipt of a liver transplant. Recent studies evaluated the MELD score in different clinical situations with positive results in patients with heart failure [
35]. In the present study, the MELD score revealed no differences between groups A and C, and the multivariate analysis showed no significant impact of the MELD score on TGC
Cmax. Hence, the MELD score appears not to be a reliable diagnostic tool to quantify liver dysfunction in critically ill patients.
In the present study, BMI was one parameter showing promising differences between the study groups at baseline. Patients of group A (high
Cmax) revealed a significantly lower mean BMI (26.4 kg/m
2) compared to patients of group C (low
Cmax, BMI 31.3 kg/m
2). The low protein binding and high distribution of TGC may be one possible explanation of this effect. However, in the multivariate analysis, BMI failed to qualify as a predictor of TGC variability. These findings are in concert with the results of other authors. Xie et al. found in their study, that BMI was an important parameter influencing the total CL of TGC. The authors pointed out that in their model building process, the simulations were beyond the BMI of the patients included and should be considered cautiously [
36]. On the other hand, Pai et al. characterized the concentration profiles of TGC in the serum and urine of obese and normal-weight healthy adults and found no differences [
37]. In accordance with these data, BMI may provide as one parameter influencing TGC PK, but the exclusive impact on TGC distribution appears weak.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.