Critically ill patients are at high risk for development of life-threatening infection leading to sepsis and multiple organ failure. Adequate antimicrobial therapy is pivotal for optimizing the chances of survival. However, efficient dosing is problematic because pathophysiological changes associated with critical illness impact on pharmacokinetics of mainly hydrophilic antimicrobials. Concentrations of hydrophilic antimicrobials may be increased because of decreased renal clearance due to acute kidney injury. Alternatively, antimicrobial concentrations may be decreased because of increased volume of distribution and augmented renal clearance provoked by systemic inflammatory response syndrome, capillary leak, decreased protein binding and administration of intravenous fluids and inotropes. Often multiple conditions that may influence pharmacokinetics are present at the same time thereby excessively complicating the prediction of adequate concentrations. In general, conditions leading to underdosing are predominant. Yet, since prediction of serum concentrations remains difficult, therapeutic drug monitoring for individual fine-tuning of antimicrobial therapy seems the way forward.

Controversy reigns as to how protein binding changes alter the time course of unbound drug concentrations in patients. Given that the unbound concentration is responsible for drug efficacy and potential drug toxicity, this area is of significant interest to clinicians and academics worldwide. The present uncertainty means that many questions relating to this area exist, including “How important is protein binding?”, “Is protein binding always constant?”, “Do pH and temperature changes alter binding?” and “How do protein binding changes affect dosing requirements?”. In this paper, we seek to address these questions and consider the data associated with altered pharmacokinetics in the presence of changes in protein binding and the clinical consequences that these may have on therapy, using examples from the critical care area. The published literature consistently indicates that a change in the protein binding and unbound concentrations of some drugs are common in certain specific patient groups such as the critically ill. Changes in pharmacokinetic parameters, including clearance and apparent volume of distribution (V_d), may be dramatic. Drugs with high protein binding, high intrinsic clearance (e.g. clearance by glomerular filtration) and where dosing is not titrated to effect are most likely to be affected in a clinical context. Drugs such as highly protein bound antibacterials with multiple half-lives within a dosing interval and that have some level of renal clearance, such as ertapenem, teicoplanin, ceftriaxone and flucloxacillin, are commonly affected. In response to these challenges, clinicians need to adapt dosing regimens rationally based on the pharmacokinetic/pharmacodynamic characteristics of the drug. We propose that further pharmacokinetic modelling-based research is required to enable the design of robust dosing regimens for drugs affected by altered protein binding.

Prolonged treatment with linezolid may cause toxicity. The purpose of this study was to define pharmacodynamic thresholds for improving safety outcomes of linezolid. We performed a retrospective study of patients who had trough (C_min) and peak (C_max) plasma levels measured during prolonged linezolid treatment. Dosage adjustments were performed when C_min ≥10 mg/L and/or AUC₂₄ ≥400 mg/L · h. Patients were divided into two subgroups according to the absence or presence of co-treatment with rifampicin (the linezolid group and the linezolid + rifampicin group, respectively). Data on demographic characteristics, disease, microbiology and haematochemical parameters were collected and outcomes in relation to drug exposure were compared between groups. A total of 45 patients were included. Dosage adjustments were needed in 40% versus 0% of patients in the linezolid group (n = 35) versus the linezolid + rifampicin group (n = 10), respectively. Patients in the linezolid group had either significantly higher C_min [3.71 mg/L (1.43–6.38) versus 1.37 mg/L (0.67–2.55), P < 0.001] or AUC₂₄ [212.77 mg/L · h (166.67–278.42) versus 123.33 mg/L · h (97.36–187.94), P < 0.001]. Thrombocytopenia appeared in 51.4% versus 0% of cases in the linezolid group versus the linezolid + rifampicin group, respectively. In 33.3% of those patients who were experiencing thrombocytopenia, therapeutic drug monitoring (TDM)-guided dosage reductions allowed recovery from toxicity and prosecution of therapy with good outcome. A logistic regression model for thrombocytopenia estimated a probability of 50% in the presence of C_min of 6.53 mg/L and/or of AUC₂₄ of 280.74 mg/L · h. Maintenance over time of C_min between 2 and 7 mg/L and/or of AUC₂₄ between 160 and 300 mg/L · h may be helpful in improving safety outcomes while retaining appropriate efficacy in adult patients receiving prolonged linezolid treatment.

Emerging evidence supports the use of therapeutic drug monitoring (TDM) of β-lactams for intensive care unit (ICU) patients to optimize drug exposure, although limited detail is available on how sites run this service in practice. This multicentre survey study was performed to describe the various approaches used for β-lactam TDM in ICUs. A questionnaire survey was developed to describe various aspects relating to the conduct of β-lactam TDM in an ICU setting. Data sought included: β-lactams chosen for TDM, inclusion criteria for selecting patients, blood sampling strategy, analytical methods, pharmacokinetic (PK)/pharmacodynamic (PD) targets and dose adjustment strategies. Nine ICUs were included in this survey. Respondents were either ICU or infectious disease physicians, pharmacists or clinical pharmacologists. Piperacillin (co-formulated with tazobactam) and meropenem (100% of units surveyed) were the β-lactams most commonly subject to TDM, followed by ceftazidime (78%), ceftriaxone (43%) and cefazolin (43%). Different chromatographic and microbiological methods were used for assay of β-lactam concentrations in blood and other biological fluids (e.g. CSF). There was significant variation in the PK/PD targets (100% fT_>MIC up to 100% fT_>4×MIC) and dose adjustment strategies used by each of the sites. Large variations were found in the type of β-lactams tested, the patients selected for TDM and drug assay methods. Significant variation observed in the PK/PD targets and dose adjustment strategies used supports the need for further studies that robustly define PK/PD targets for ICU patients to ensure a greater consistency of practice for dose adjustment strategies for optimizing β-lactam dosing with TDM.