Antimicrobial stewardship

Contents

Overview

Antimicrobial stewardship (AMS) encompasses coordinated interventions to improve the appropriateness of antimicrobial prescribing. In the ICU, broad-spectrum empirical therapy is often necessary but must be followed by systematic review, de-escalation, and timely cessation. The ICU is a major driver of resistance selection, and stewardship is a patient safety and public health imperative.

Principles of Stewardship

The overarching aim is the right drug, at the right dose, via the right route, for the right duration.

Empirical therapy: Broad-spectrum agents are appropriate at initiation in sepsis, before microbiological results are available. Antibiotic choice should be guided by the likely site of infection, local resistance patterns, host immune status, and prior antibiotic exposure.

48–72 hour review: Once culture results, sensitivities, and clinical response are available, each antibiotic course must be formally reviewed. Three outcomes are possible: de-escalate, continue unchanged, or stop.

De-escalation: Narrowing the spectrum of antibiotic therapy once the causative organism and its sensitivities are known. For example, switching from piperacillin-tazobactam to co-amoxiclav if a sensitive organism is identified, or from vancomycin to flucloxacillin once MRSA is excluded. De-escalation is safe and reduces selection pressure for resistance without worsening patient outcomes.

Duration: The shortest effective course is preferred. Procalcitonin-guided antibiotic discontinuation reduces antibiotic days in ICU without increasing mortality or recurrence, and is recommended in current guidelines.

IV to oral switch: Conversion to oral antibiotics is safe once the patient is absorbing enterally, the clinical condition is improving, and appropriate oral agents are available. Reduces IV line complications and cost.

Pharmacokinetics and Pharmacodynamics

Understanding antibiotic PK/PD is essential for effective dosing, particularly in critically ill patients where pharmacokinetics are profoundly altered.

PK/PD Parameters

Time-dependent antibiotics: Efficacy depends on the proportion of time the free drug concentration exceeds the minimum inhibitory concentration (T>MIC). The goal is to maximise the duration of exposure.

Examples: beta-lactams (penicillins, cephalosporins, carbapenems), vancomycin (AUC/MIC), clindamycin.

Strategy: Extended or continuous infusion of beta-lactams increases T>MIC significantly compared with intermittent bolus dosing. For example, meropenem 1 g over 3 hours is superior to 1 g over 30 minutes for organisms at the upper range of sensitivity.

Concentration-dependent antibiotics: Efficacy depends on the peak drug concentration relative to the MIC (Cmax/MIC). The goal is a high peak.

Examples: aminoglycosides, metronidazole, fluoroquinolones (also have a significant AUC/MIC component).

Strategy: Once-daily aminoglycoside dosing achieves a higher peak and also exploits the post-antibiotic effect, while limiting nephrotoxicity by allowing a drug-free trough period.

Therapeutic Drug Monitoring

Vancomycin: AUC/MIC-guided dosing is now preferred over trough-guided monitoring. An AUC/MIC of 400–600 mg·h/L is associated with optimal efficacy and reduced nephrotoxicity.

Aminoglycosides: Extended-interval dosing with trough monitoring (target <1 mg/L for gentamicin). Single-sample Bayesian models improve precision.

Beta-lactams: TDM is available in specialised centres; particularly valuable in augmented renal clearance or resistant organisms.

Resistance Mechanisms

Beta-lactamase production:

  • Extended-spectrum beta-lactamases (ESBLs): produced by Enterobacterales (E. coli, Klebsiella); hydrolyse most penicillins and cephalosporins. Require carbapenem or tazobactam-containing agents.
  • Carbapenemases: KPC (Klebsiella pneumoniae carbapenemase), NDM (New Delhi metallo-beta-lactamase), OXA-48; hydrolyse carbapenems. Associated with very limited treatment options; require ceftazidime-avibactam, meropenem-vaborbactam, or imipenem-relebactam.

Efflux pumps: Active export of antibiotics from the cell. Important in Pseudomonas aeruginosa and Acinetobacter baumannii.

Target modification: Altered penicillin-binding proteins (PBP2a in MRSA, encoded by mecA gene); ribosomal methylation conferring macrolide and aminoglycoside resistance.

Outer membrane porin loss: Reduced permeability to carbapenem entry; particularly important in Pseudomonas and Acinetobacter.

Multiple mechanisms may coexist, producing extensively drug-resistant (XDR) or pan-drug-resistant (PDR) organisms.

Key Stewardship Interventions

Antibiotic time-out: Formal review at 48–72 hours to reassess the indication, spectrum, dose, and duration. Should be embedded in daily ICU ward round documentation.

Prospective audit and feedback: A pharmacy or infectious disease clinician reviews antibiotic prescribing and provides direct feedback to prescribers. Highly effective at reducing inappropriate prescribing.

Procalcitonin guidance: Serial procalcitonin measurement guides antibiotic cessation. PCT-guided protocols reduce antibiotic duration in VAP and other ICU infections without increasing mortality.

Blood culture timing: Two sets of blood cultures should be taken before antibiotics whenever possible. Each hour of delay to antibiotics in septic shock increases mortality — but the 10 minutes required to obtain cultures before starting antibiotics does not.

Biomarker-guided prescribing: CRP, white cell count, and clinical trajectory all inform the decision to de-escalate or stop, particularly when cultures are negative.

ICU-Specific Challenges

Augmented renal clearance (ARC): A GFR significantly above normal (>130 mL/min/1.73 m²), common in young patients with sepsis and multiple organ support. Renally cleared antibiotics are eliminated more rapidly, resulting in subtherapeutic levels at standard doses. Beta-lactams and vancomycin are particularly affected. ARC can be identified with a 24-hour urinary creatinine clearance measurement or estimated using the CG-ARC score. Higher and more frequent doses, or continuous infusion, may be required.

Altered volume of distribution: Fluid resuscitation and third spacing increase Vd for hydrophilic antibiotics, reducing peak concentrations. This is particularly relevant for beta-lactams and aminoglycosides at the start of critical illness.

ECMO: Drug sequestration in the ECMO circuit significantly reduces concentrations of many antibiotics, particularly lipophilic agents (voriconazole) and some beta-lactams. TDM is essential in patients on ECMO.

Frequent culture-negative infection: Many patients receive prolonged empirical antibiotics for suspected VAP or line infections without positive cultures. Stopping antibiotics in clinically improving culture-negative patients is supported by evidence and reduces harm.

Viva Questions

What is the difference between time-dependent and concentration-dependent antibiotics? Give examples.

Time-dependent antibiotics exert their bactericidal effect when drug concentrations remain above the minimum inhibitory concentration for as long as possible during the dosing interval. The key PK/PD parameter is time above MIC (T>MIC). Beta-lactams — penicillins, cephalosporins, and carbapenems — are the principal examples. They display a flat dose-response relationship above the MIC, so increasing the dose above the MIC threshold does not improve efficacy; instead, extending the duration of exposure does. This is the rationale for extended or continuous infusion of meropenem or piperacillin-tazobactam in the ICU. Concentration-dependent antibiotics achieve maximum bactericidal activity when peak drug concentration is as high as possible relative to the MIC. The key parameter is Cmax/MIC. Aminoglycosides are the classic example; once-daily dosing achieves a high peak, exploits the post-antibiotic effect, and allows a low-trough period that reduces nephrotoxicity. Fluoroquinolones are both concentration-dependent and exhibit AUC/MIC-dependent killing.

How would you approach de-escalation in a ventilated patient with improving sepsis?

At the 48–72 hour mark I would formally review the antimicrobial regimen as part of the daily ward round. I would assess the clinical trajectory — improvement in vasopressor requirement, temperature, white cell count, and CRP — and review all microbiological results: blood cultures, respiratory specimens if VAP was suspected, and any other site-specific cultures. If cultures have identified the causative organism and sensitivities, I would switch to the narrowest appropriate agent. If cultures are negative and the clinical picture is improving, I would strongly consider stopping antibiotics after an appropriate empirical course — typically 5–7 days for VAP or bacteraemia, guided by procalcitonin if available. I would document the decision and the rationale in the notes. De-escalation should not wait for perfect certainty; a clinically improving patient with negative cultures has a low probability of ongoing bacterial infection requiring broad-spectrum coverage. Continuing broad-spectrum antibiotics beyond clinical resolution increases the risk of C. difficile, resistant organism selection, and drug toxicity.

What is augmented renal clearance and why does it matter for antibiotic dosing?

Augmented renal clearance describes a state where the glomerular filtration rate is significantly above normal — typically defined as a creatinine clearance exceeding 130 mL/min/1.73 m². It is most common in young, previously healthy patients with sepsis, trauma, or haematological malignancy, particularly during the early inflammatory phase of critical illness. In these patients, standard doses of renally cleared antibiotics — beta-lactams, vancomycin, aminoglycosides, piperacillin-tazobactam — are eliminated far more rapidly than expected, producing subtherapeutic plasma and tissue concentrations that may result in treatment failure or resistance selection. Standard eGFR equations underestimate renal function in these patients; a 24-hour urinary creatinine clearance is the most reliable measure. Management involves increasing dose frequency, using higher absolute doses, and for time-dependent agents considering continuous infusion combined with therapeutic drug monitoring to confirm adequate exposure.