Antimicrobial stewardship & spectrum

What stewardship is and why it matters

Antimicrobial stewardship is the coordinated set of interventions designed to optimize antibiotic use. It is not about denying antibiotics — it is about giving the right drug, at the right dose, by the right route, for the right duration, and only when indicated. The benefits accrue to the individual patient and to the population.

• Better outcomes: appropriate therapy faster, fewer adverse events.
• Less resistance: antibiotic pressure drives resistance at the unit and hospital level.
• Less C. difficile: antibiotic exposure is the single biggest modifiable risk factor.
• Lower cost and toxicity: unnecessary days and redundant coverage add up.

The antibiogram & the empiric choice

Empiric therapy is chosen before susceptibilities are known, so it must be informed by the likely organism and local resistance patterns. The hospital antibiogram summarizes the susceptibility of common isolates over the prior year and is the indispensable reference for empiric decisions at your institution.

• Identify the syndrome and the likely organisms first; then pick a drug that covers them.
• Cross-check against your unit antibiogram — empiric choices that worked elsewhere may fail locally.
• Weigh host factors: allergy, renal/hepatic function, weight, pregnancy, immunocompromise, recent antibiotics.
• Avoid redundancy: double anaerobic or double MRSA coverage is rarely needed empirically.

De-escalation, narrowing, and the antibiotic timeout

The most impactful single stewardship habit is the structured review of every antibiotic at 48–72 hours, once culture and susceptibility data return. This 'timeout' asks three questions: is the antibiotic still needed, can it be narrowed, and how long should it run?

• Stop empiric coverage for organisms not isolated (e.g., stop MRSA coverage when cultures grow an organism without MRSA risk).
• Narrow to the narrowest effective agent based on susceptibilities.
• Convert IV to oral when the patient is stable and an oral option with good bioavailability exists.
• Set a clear stop date or reassessment date at the time of ordering.

Duration: less is more

For decades, courses were long by tradition. Modern trials have repeatedly shown that shorter durations are as effective for many common infections and carry less risk of resistance and C. difficile. The 2022 BACTEREMIA individual-patient-data meta-analysis of 9 randomized trials and the PITT trial (von Dach et al. 2023) established that 7 days is non-inferior to 14 days for uncomplicated Enterobacterales bacteremia in clinically responding patients. When a patient is improving and source control is adequate, default to the shortest evidence-based duration.

• Uncomplicated cystitis: 3–5 days (nitrofurantoin 5 days, TMP-SMX 3 days, fosfomycin single-dose).
• Pyelonephritis: 5–7 days (fluoroquinolone) or 7–14 days (beta-lactam).
• Community-acquired pneumonia (responding): ≥5 days minimum, if afebrile 48–72 h and clinically stable.
• Cellulitis (non-purulent): 5–6 days.
• Uncomplicated gram-negative bacteremia: 7 days from first negative culture (PITT; BACTEREMIA meta-analysis — non-inferior to 14 days). [VERIFY PITT inclusion criteria and 7-day definition]
• Intra-abdominal infection (source-controlled): 4 days post-source-control (STOP-IT).
• Do NOT shorten when: source control incomplete, endovascular infection, slow clinical response, metastatic foci, or immunocompromise.

PK/PD: time- vs concentration-dependent, and why it changes dosing

Pharmacokinetic/pharmacodynamic (PK/PD) principles describe how a drug's exposure relates to its effect and guide dosing and optimization. Beta-lactams kill based on the fraction of the dosing interval free drug concentrations exceed the organism's MIC (fT>MIC); aminoglycosides, fluoroquinolones, and daptomycin kill based on peak exposure (Cmax/MIC or AUC/MIC). Translating PK/PD to the bedside is the foundation of modern optimization — extended infusions, once-daily aminoglycosides, and AUC-guided vancomycin.

• Beta-lactams: maximize fT>MIC — for severe pseudomonal or resistant infection use extended (3–4 h) or continuous infusion of piperacillin-tazobactam, cefepime, or meropenem. [VERIFY outcome evidence]
• Aminoglycosides: extended-interval (once-daily) dosing exploits concentration-dependent killing and a long post-antibiotic effect; monitor for nephro/ototoxicity.
• Vancomycin: AUC₀₋₂₄ 400–600 for serious MRSA infection — Bayesian dosing (2 timed levels or software) is preferred over the 2-trapezoidal-level method; trough-only monitoring is no longer recommended (Rybak 2020 consensus). [VERIFY Bayesian vs trapezoidal recommendation strength]
• Daptomycin: concentration-dependent; dosing escalates with weight and indication (6 mg/kg for S. aureus bacteremia; higher for VRE).
• Vancomycin + piperacillin-tazobactam: associated with higher AKI than either alone — reconsider the combination when both are not essential. [VERIFY magnitude of AKI risk]

IV-to-oral conversion: bioavailability drives the step-down

Early IV-to-oral switch shortens length of stay, line days, and cost without harming outcomes — provided the patient is hemodynamically stable, tolerating oral intake, and the chosen oral agent reaches the infection site. The decision rests almost entirely on oral bioavailability: drugs with near-complete bioavailability are pharmacokinetically equivalent IV and PO, so continuing IV therapy is purely a delivery issue, not an efficacy one.

• Near-100% bioavailable (effectively interchangeable IV/PO): fluoroquinolones, linezolid, metronidazole, TMP-SMX, fluconazole, and doxycycline.
• Good but variable bioavailability: beta-lactams (amoxicillin ~80%, cephalexin ~90%, cefuroxime ~50%), clindamycin (~90%), and azithromycin (~37%).
• Convert when: clinically improving, afebrile 24–48 h, tolerating PO, hemodynamically stable, and an oral agent with adequate bioavailability and tissue penetration exists.
• Avoid oral step-down when absorption is unreliable (ileus, neutropenic enterocolitis, severe shock, malabsorption) or the source is endovascular, CNS, or undrained.

Penicillin allergy delabeling: a high-yield stewardship intervention

Roughly 9 in 10 patients who carry a penicillin allergy label are not truly allergic, yet the label drives broader, more toxic alternatives (vancomycin, fluoroquinolones, aztreonam) and is associated with longer stays, more resistance, and worse outcomes. Delabeling is a core stewardship action and can be done safely at the bedside with a validated tool. The PEN-FAST score combines a brief history (PENicillin allergy, New reaction, Severe reaction, Time since last reaction) with a point score that stratifies risk and identifies candidates for direct oral amoxicillin challenge without skin testing.

• PEN-FAST 0: very low risk — oral amoxicillin challenge (single dose, observed) without prior skin testing. [VERIFY exact PEN-FAST components and cutoffs]
• PEN-FAST 1–2: low-moderate risk — skin testing or graded oral challenge under supervision.
• PEN-FAST ≥3: higher risk — referral to allergy/immunology; consider skin testing or supervised challenge.
• Side-chain cross-reactivity (not the beta-lactam ring) is what matters: cefazolin and ceftriaxone do not share side chains with penicillin and are safe in most penicillin-allergic patients without anaphylaxis.
• Anaphylaxis, SJS/TEN, or severe delayed reactions (DRESS, AGEP) are NOT candidates for bedside delabeling — refer to allergy.

Adverse effects & toxicity

Every antibiotic has a toxicity profile. Knowing the signature adverse effects lets you anticipate, monitor, and choose safer alternatives — and it is high-yield for both the wards and the boards.

• Beta-lactams: allergy, interstitial nephritis, C. difficile.
• Vancomycin: nephrotoxicity (especially with piperacillin-tazobactam), infusion reaction, ototoxicity at high levels.
• Aminoglycosides: nephrotoxicity, ototoxicity — monitor levels and duration.
• Fluoroquinolones: QT prolongation, tendinopathy/rupture, dysglycemia, CNS effects, aortic warning — reserve for clear indications.
• Daptomycin: inactivated by pulmonary surfactant — never for pneumonia; monitor creatine kinase for myopathy.
• Linezolid: thrombocytopenia, neuropathy, and serotonin syndrome with prolonged use (>2 weeks) or serotonergic drugs.
• Colistin/polymyxin: nephrotoxicity and neurotoxicity — drug of last resort.

Special situations: asymptomatic bacteriuria & prophylaxis

Two scenarios generate the most unnecessary antibiotics on a medicine service: treating asymptomatic bacteriuria and prolonging surgical prophylaxis. Both are teachable moments. Treating asymptomatic bacteriuria does not prevent symptomatic infection in most patients and increases harm — with two well-defined exceptions.

• Do NOT treat asymptomatic bacteriuria — except in pregnancy and before urologic procedures with mucosal bleeding.
• Pyuria alone does not require treatment in the absence of symptoms.
• Surgical prophylaxis: single preoperative dose; stop within 24 hours (48 hours for cardiac); do not continue as 'coverage.'
• Treat the patient, not the culture or the colonization.

Guidelines

• Implementing an Antibiotic Stewardship Program (IDSA/SHEA)Barlam TF, et al. Clin Infect Dis. 2016;62(10):e51–e77The core guideline for building and running stewardship interventions.
• A consensus guideline for therapeutic drug monitoring of vancomycin (AUC-based)Rybak MJ, et al. Am J Health-Syst Pharm. 2020Establishes AUC-based vancomycin monitoring (target AUC 400–600).
• Clinical Practice Guideline for Clostridioides difficile Infection (IDSA/SHEA)McDonald LC, et al. Clin Infect Dis. 2018 [updated 2021; fidaxomicin preferred] [VERIFY current version]Fidaxomicin first-line for initial and recurrent CDI; oral vancomycin alternative; bezlotoxumab for recurrence risk.
• 2024 IDSA Guidance on the Treatment of Antimicrobial-Resistant Gram-Negative Infections (AMR)Tamma PD, et al. Clin Infect Dis. 2024 [VERIFY]Carbapenem-sparing and targeted therapy for ESBL, CRE, and DTR Pseudomonas — including novel beta-lactamase inhibitors.

Reviews

• The New Antibiotic Mantra — “Shorter Is Better”Spellberg B. N Engl J MedA persuasive overview of why shorter durations matter.
• Risk stratification of penicillin allergy by clinical history (PEN-FAST review)Stone CA Jr, et al. JAMA. 2020 [VERIFY]Reviews PEN-FAST application, cross-reactivity, and the safety of direct oral challenge in low-risk patients.

Key trials

• Short versus long duration of therapy for uncomplicated gram-negative bacteremia (PITT)von Dach E, et al. JAMA Intern Med. 2023A key trial supporting 7-day courses for uncomplicated gram-negative bacteremia.
• MERINO — Piperacillin-tazobactam vs ceftriaxone for ESBL E. coli / K. pneumoniae bacteremiaHarris PNA, et al. JAMA. 2018;320(10):984–994Piperacillin-tazobactam was inferior to carbapenems for ESBL bloodstream infection — use a carbapenem. [VERIFY volume/pages]
• Seven versus 14 days of antibiotic therapy for uncomplicated gram-negative bacteremia (individual patient data meta-analysis)Yahav D, et al. JAMA. 2022;327(4):335–345 [VERIFY — 9 RCTs IPD meta]7 days non-inferior to 14 days for uncomplicated Enterobacterales bacteremia across multiple trials.
• PEN-FAST — A tool for rapid risk stratification of penicillin allergyDevchand M, et al. J Antimicrob Chemother. 2019;74(2):437–442 [VERIFY exact volume/pages]Validated 4-item score to identify low-risk penicillin-allergy labels for direct oral challenge without skin testing.