Introduction
Effective treatment of sepsis demands not only the prompt administration of appropriate antimicrobials but also precise dosing to enhance patient survival. Adequate dosing ensures therapeutic drug concentrations at the infection site, leading to favorable clinical and microbiological outcomes while minimizing antibiotic-related toxicity. Therapeutic drug monitoring (TDM) is the recommended approach for achieving these goals, but it is not universally available in all intensive care units (ICUs) or for all antimicrobials. In such cases, clinicians must consider multiple factors, such as the patient’s clinical condition, causative pathogen, organ dysfunction, and the physicochemical properties of the drugs. The pharmacokinetics (PK) of antimicrobials can vary significantly among critically ill patients and even within the same patient during their ICU stay, underscoring the need for individualized dosing strategies.
Critically ill patients exhibit unique pathophysiological changes that significantly influence the PK and pharmacodynamics (PD) of antibiotics, posing challenges to effective infection management. Altered organ function, tissue perfusion changes, and fluid balance variations can impact drug absorption, distribution, metabolism, and elimination. Standard dosing regimens often fail to achieve therapeutic concentrations in such patients, potentially leading to suboptimal outcomes or resistance. Optimizing antibiotic therapy requires a deep understanding of drug-pathogen interactions, including the antibiotic’s mechanism of action, the pathogen’s minimal inhibitory concentration (MIC), and the exposure duration needed for bactericidal effects. This review highlights the critical factors affecting antibiotic PK and PD in critically ill patients and offers practical recommendations for clinicians to optimize dosing, ensuring effective antimicrobial therapy and improved patient outcomes.
Pathophysiological Alterations in Intensive Care Unit (ICU) Patients
Drug PK may be significantly affected in the presence of pathophysiological changes that occur during critical illness. In conjunction with the strategies used for the early management of critical illness (such as administration of fluids and vasopressors), the development of systemic inflammatory response syndrome (in the setting of major surgery, trauma, burns, or sepsis) significantly affects the two major PK parameters related to drug dosing, namely, the volume of distribution (Vd) and drug clearance (CL).
Changes in Vd
Changes in Vd in Sepsis
Sepsis-induced vascular permeability and edema cause fluid shifts from the intravascular to the interstitial space, increasing the volume of distribution (Vd) of antimicrobial drugs. Fluid resuscitation and the use of inotropes and vasopressors during initial management exacerbate this effect. Hydrophilic drugs like β-lactams, aminoglycosides, and vancomycin with low Vd are particularly impacted. Conditions like pleural effusion, ascites, and surgical drains further expand the Vd.
Hypoalbuminemia, common in critically ill patients, reduces drug-albumin binding for highly protein-bound drugs such as ceftriaxone, ertapenem, and teicoplanin, increasing the free drug fraction and further expanding Vd. This leads to lower peak plasma drug concentrations (Cmax) and potential underdosing. Dose adjustments, especially in the initial treatment phase, are crucial to maintain therapeutic drug levels.
Changes in drug CL
The progression of critical illnesses, such as sepsis, often leads to multiple organ dysfunction syndrome (MODS), impacting drug pharmacokinetics (PK). Impaired perfusion of organs like the gastrointestinal tract and kidneys can reduce drug absorption and elimination. Reduced hepatic blood flow further diminishes the metabolism of drugs with a high hepatic extraction ratio, while chronic liver failure may necessitate dose adjustments for drugs like echinocandins. Augmented renal clearance (ARC), a phenomenon in critically ill patients with hyperdynamic states, enhances renal elimination and significantly affects hydrophilic drugs like β-lactams, vancomycin, and aminoglycosides. Studies show β-lactam underexposure in ARC, requiring high-dose regimens to improve outcomes without increasing adverse effects.
Renal impairment and the need for renal replacement therapy (RRT) add complexity to drug dosing in critically ill patients. More than 50% of ICU patients experience acute kidney injury, with 20%–25% requiring RRT. Drug PK during RRT depends on extracorporeal therapy characteristics, such as membrane permeability and drug properties like molecular size and protein binding. Variability in antibiotic concentrations due to unadjusted dosing risks toxicity or insufficient therapeutic levels. Tailored drug regimens, daily creatinine clearance monitoring, and consideration of alternatives with low renal clearance are essential to optimize treatment and minimize risks of inappropriate dosing in these patients.
Changes in drug absorption
Altered drug absorption has been previously reported in critically ill patients. However, there are no clear recommendations for the management of those alterations. Critical illnesses may affect the gastrointestinal tract and lead to a decrease in intestinal peristalsis, mucosal impairment, and altered drug metabolism. Enteric drug absorption and availability are difficult to predict, mainly due to fluctuations in gastric pH, loss of enteric architecture, and decreased enzymatic activity. In addition, the delay in gastric emptying extends the time needed to achieve maximum concentrations of the antibiotic. The impact of these pathophysiological alterations is illustrated by the significant decrease in the absorption of antibiotics such as ciprofloxacin in ICU patients.
Changes in tissue penetration
The transport of antibiotics to tissues and subsequent distribution within tissues and cells depends on various factors including the characteristics of the drug itself, patient characteristics (e.g., obesity), disease severity, and target tissues.
Notably, different antimicrobial drugs demonstrate altered tissue penetration in critically ill patients. As reported in the literature, the extent of these alterations varies widely among different tissues and organs. In the presence of sepsis, the microcirculatory blood flow may be significantly impaired due to endothelial dysfunction and the presence of microthrombi, which decrease tissue perfusion. These alterations may lead to suboptimal antibiotic exposure at the site of infection and thereby to potential therapeutic failure, emergence of resistance, and higher morbidity. Suboptimal tissue concentrations may even be found in patients with adequate plasma concentrations, as the antibiotic concentrations in plasma do not accurately reflect those in infected tissue. Interestingly, clinical scoring systems such as the tissue penetration prediction score have been proposed for predicting tissue penetration of antimicrobials. The main factors found to correlate with tissue penetration include oxygen saturation, serum lactate levels, and the dose per time unit of norepinephrine. Although the tissue/plasma penetration ratio of antimicrobials may be an important factor for the selection of the most suitable treatment, there is currently no conclusive evidence to support the use of clinical scores for adjusting antibiotic dosing regimens. In addition, the limited availability of data pertaining to tissue penetration precludes their use in guiding antimicrobial dosing.
Changes due to extracorporeal therapies
Extracorporeal membrane oxygenation (ECMO) is a life-support system used for patients with severe respiratory or cardiac failure, providing cardiopulmonary support and serving as a bridge to recovery or transplantation. The ECMO circuit, consisting of a pump, oxygenator, heat exchanger, and tubing, can sequester drugs, alter apparent volume of distribution (Vd), and reduce drug clearance (CL). Drug sequestration depends on properties like lipophilicity and protein binding, as well as circuit design. Hemodilution from the circuit’s priming solution impacts drugs with low Vd, such as β-lactams, more significantly than those with high Vd, like quinolones.
Drug clearance during ECMO is often reduced due to impaired renal and hepatic perfusion, and the simultaneous use of renal replacement therapy (RRT) in nearly half of ECMO patients adds further complexity. Dual extracorporeal circuits complicate pharmacokinetic (PK) parameter estimation, necessitating integrated approaches combining ex vivo experiments, animal models, and clinical studies to optimize drug dosing. Advances in circuit technology aim to minimize these PK challenges.
Applying PK/PD Approaches to Optimize Antimicrobial Therapy
PK/PD targets for critically ill patients
Pharmacokinetic/pharmacodynamic (PK/PD) targets for antibiotics vary based on killing mechanisms: dose-dependent (Cmax/MIC), time-dependent (%T > MIC), or AUC-dependent (AUC/MIC). For β-lactams, a target of 100% fT > MIC is often suggested, with higher targets (4–5× MIC) considered to reduce resistance and microbiological failure. Studies link suboptimal Cmin/MIC ratios (e.g., ≤5) to poor outcomes in infections like VAP caused by Gram-negative bacteria. While observational studies support higher PK/PD targets for better remission, their impact on mortality is unclear.
In early infections, when actual MICs are unavailable, epidemiological cut-off (ECOFF) values based on local ecology are used, though they may underestimate target attainment. Variability in MIC measurements and delays in results complicate therapy optimization. Rapid diagnostic tools and susceptibility testing could help tailor antimicrobial therapy more effectively in the future.
Increased loading doses
Timely antibiotic administration is critical in sepsis, but standard dosing often results in subtherapeutic levels. Increased Vd in sepsis necessitates higher loading doses for hydrophilic antibiotics like β-lactams, vancomycin, aminoglycosides, and colistin. New dosing strategies, such as 8 g/3 h for piperacillin or 2 g/0.5 h for meropenem, are derived from population PK studies and simulations, improving PK/PD target attainment.
Clinical validation supports higher loading doses for these antibiotics, even in renal impairment or during RRT, as loading doses remain unaffected by renal function. Tailored dosing nomograms, considering factors like body weight and renal clearance, could further optimize initial antibiotic therapy.
Optimal mode of administration of antibiotics
Continuous Infusion Mode
Traditionally, beta-lactam antibiotics have been administered via intermittent infusions. However, research increasingly supports continuous infusion in specific clinical scenarios. Studies suggest continuous administration improves pharmacokinetic/pharmacodynamic (PK/PD) target achievement, clinical remission rates, and microbiological eradication. However, meta-analyses of randomized controlled trials (RCTs) have not consistently shown a survival advantage, likely due to small sample sizes and underpowered studies. A recent meta-analysis of individual patient data from multicenter RCTs found lower hospital mortality with continuous infusion compared to intermittent infusion (19.6% vs. 26.3%; relative risk = 0.74, P = 0.045). Conversely, a large RCT of meropenem in critically ill patients showed no significant difference in mortality or resistance emergence. Ongoing trials, like BLING III, are expected to provide more clarity on the impact of continuous infusion. Observational studies on non-beta-lactam antibiotics like linezolid have demonstrated improved PK/PD target achievement, alveolar diffusion, and clinical outcomes with continuous infusion. However, RCTs comparing continuous and intermittent infusions remain limited.
Inhaled Antibiotics
Inhaled antibiotics offer a targeted approach to treating pulmonary infections while minimizing multidrug-resistant (MDR) strain emergence. Effective lung deposition of high doses is critical, necessitating the use of mesh nebulizers and optimized ventilator settings. Meta-analyses indicate higher clinical recovery and microbiological eradication rates with inhaled antibiotics, but no significant mortality or nephrotoxicity benefits. Recent trials like INHALE and VAPORISE assessed inhaled aminoglycosides and fosfomycin as adjunctive therapies for ventilator-associated pneumonia (VAP) caused by Gram-negative bacteria but found no advantage over standard intravenous treatment. Meanwhile, a prophylactic approach using inhaled amikacin demonstrated a significant reduction in VAP development in mechanically ventilated patients. However, non-inferiority trials comparing inhaled polymyxins to newer intravenous antibiotics for MDR infections are still needed.
Therapeutic Drug Monitoring (TDM) and Dosing Software
Variability in drug response among patients complicates treatment, often requiring dose adjustments to optimize outcomes. Therapeutic drug monitoring (TDM) and model-informed precision dosing (MIPD) have emerged to address this challenge. Recent reviews and meta-analyses on TDM for beta-lactam antibiotics in critically ill patients showed no significant reduction in mortality or antimicrobial resistance but highlighted biases related to non-adherence to TDM protocols. While TDM has been associated with better target achievement and reduced underdosing, its benefits in mortality and clinical outcomes remain unclear. Challenges include delays in obtaining TDM results and interpreting them, particularly in non-specialized centers. Reducing turnaround times and improving interpretation support are critical for effective implementation. The importance of minimum inhibitory concentration (MIC) selection and timely result delivery is highlighted as crucial for clinical success.
Monitoring the Side Effects of Antibiotics
Evidence underscores the potential for significant toxicity with beta-lactam antibiotics, particularly in intensive care patients. TDM is essential for mitigating toxicities related to excessive drug exposure. While higher empirical doses based on TDM do not necessarily lead to toxicity, well-established toxicity thresholds are lacking. Studies have linked elevated drug concentrations with neurotoxicity, such as cefepime levels above 22 mg/L or piperacillin levels above 360 mg/L. Thresholds for neurotoxicity vary across antibiotics and patient conditions, emphasizing the need for toxicodynamic target definitions. Ongoing trials like the OPTIMAL TDM study aim to establish these thresholds and evaluate the impact of TDM-guided dosing adjustments on toxicity prevention and clinical outcomes. Enhanced TDM practices could significantly improve safety profiles and therapeutic efficacy in critical care settings.
New and Old Antibiotics in Critically Ill Patients
Repositioning Old Antibiotics
There has been a resurgence of interest in older antibiotics as a response to the growing bacterial resistance to commonly used antibacterial drugs. Some of these old antibiotics show activity against multi-drug-resistant (MDR) bacteria, offering an alternative treatment option. However, the clinical indications and dosing recommendations outlined in the product information have not been updated and may not be suitable for critically ill patients.
Colistin
Colistin, an intravenous antibiotic, is widely used, but there is confusion due to different formulations and dosing conventions. Unlike polymyxin B, colistin is administered as colistin methanesulfonate sodium (CMS), which must be converted into colistin in the body. The time to reach therapeutic plasma colistin concentrations is longer than that of polymyxin B, and significant inter- and intra-individual variations are observed in colistin pharmacokinetics (PK). Studies have highlighted the need for individualized dosing and therapeutic drug monitoring (TDM) for critically ill patients. The international consensus guidelines recommend a CMS loading dose followed by maintenance dosing, with adjustments based on renal function. However, TDM for colistin is not widely available, and its absence could lead to under or overdosing, especially in critically ill patients.
Fosfomycin
Fosfomycin is being repurposed to combat MDR pathogens, particularly against Enterobacterales and non-fermenting Gram-negative bacteria. It is a hydrophilic drug with low protein binding and good tissue penetration. However, PK studies show significant variability in critically ill patients when standard dosing is applied. Research has identified the 24-hour area under the concentration-time curve over the MIC (AUC0–24/MIC) as the best PK/PD index for fosfomycin. Prolonged and continuous infusions have been suggested to optimize drug exposure in critically ill patients, especially those with carbapenem-resistant infections. Additionally, prolonged infusion may reduce the risk of severe hypokalemia associated with fosfomycin use.
Temocillin
Temocillin, a derivative of ticarcillin, is approved in several European countries for treating urinary tract, bloodstream, and lower respiratory tract infections. Despite limited use due to issues with clinical breakpoints and therapeutic regimens, interest in temocillin has grown due to its activity against ESBL, AmpC, and KPC-producing Enterobacterales. In critically ill patients, population PK studies show significant variability, with continuous infusion regimens providing better target attainment. However, certain infections, such as intra-abdominal infections and those with ascites, may require higher doses to meet pharmacokinetic/pharmacodynamic (PK/PD) targets. MIC determination and TDM should be considered for optimal dosing in critically ill patients.
New antibiotics
New antimicrobial drugs have been developed over the past decade to address the threat of antimicrobial resistance. These mostly include combinations of a β-lactam and a β-lactamase inhibitor.
Future Perspectives and Conclusion
The need for personalized antimicrobial dosing in critically ill patients is becoming increasingly clear due to the considerable and often unpredictable pharmacokinetic variability in this population. Although therapeutic drug monitoring (TDM) can aid in precise dose adjustments, it is not universally accessible for all antimicrobial agents. Moreover, TDM does not offer guidance for initial empirical dosing, which plays a crucial role in improving patient outcomes, such as reducing mortality in septic shock cases. As a result, clinicians face significant challenges in predicting drug distribution patterns and achieving adequate drug concentrations at infection sites. This process requires understanding the complex interactions between a patient’s physiological state, underlying diseases, the drug’s properties, and any extracorporeal treatments being used.

Figure 1
To address these challenges, innovative approaches are being explored to better tailor antimicrobial dosing. Early sepsis management now includes rapid diagnostic tools that can swiftly determine the susceptibility profile of pathogens, enabling clinicians to adjust dosing regimens for resistant strains. One promising method is model-informed precision dosing (MIPD), which uses mathematical and statistical algorithms to integrate both patient-specific data and drug concentration measurements. MIPD, though complex, can be incorporated into clinical practice with software tools, supporting clinical decision-making and improving bedside implementation. Preliminary studies have shown that MIPD can reduce the incidence of vancomycin-associated nephrotoxicity and may be cost-effective for preventing nephrotoxicity in patients with renal impairment. Moving forward, well-designed clinical trials are needed to assess the benefits of precision dosing, particularly for β-lactams.
Source: Roger, Claire. “Understanding antimicrobial pharmacokinetics in critically ill patients to optimize antimicrobial therapy: A narrative review.” Journal of Intensive Medicine (2024).