What are the appropriate empiric antibiotic choices for infections caused by Gram‑negative bacilli?

Empiric Antibiotic Selection for Gram-Negative Bacilli Infections

For severe infections caused by Gram-negative bacilli, use a carbapenem (meropenem or imipenem) as first-line therapy, with combination therapy (adding an aminoglycoside or fluoroquinolone) strongly recommended for septic shock or high-risk multidrug-resistant organisms. 1, 2

Risk Stratification Determines Antibiotic Choice

Low-Risk, Non-Severe Infections

• For uncomplicated urinary tract infections (cUTI) without septic shock: Use piperacillin-tazobactam, amoxicillin-clavulanic acid, quinolones, or cotrimoxazole 1
• For cUTI specifically: Aminoglycosides (when active in vitro) or IV fosfomycin are strongly recommended 1
• Third-generation cephalosporins (ceftriaxone, cefotaxime) remain acceptable for community-acquired infections in areas with <18% resistance rates 3

Severe Infections and Bloodstream Infections (BSI)

• For BSI with septic shock due to third-generation cephalosporin-resistant Enterobacterales (3GCephRE): Carbapenems (imipenem or meropenem) are the gold standard 1
• For BSI without septic shock: Ertapenem may be substituted for imipenem/meropenem 1
• For severe intra-abdominal infections: Use cefotaxime or ceftriaxone plus metronidazole 4

Combination Therapy: When and Why

Combination therapy with a beta-lactam plus aminoglycoside or fluoroquinolone significantly reduces inappropriate initial therapy and improves mortality in severe sepsis. 2

Evidence for Combination Therapy

• Adding an aminoglycoside to carbapenem increases appropriate coverage from 89.7% to 94.2% 2
• Adding aminoglycoside to cefepime increases coverage from 83.4% to 89.9% 2
• Adding aminoglycoside to piperacillin-tazobactam increases coverage from 79.6% to 91.4% 2
• Aminoglycosides provide broader coverage than fluoroquinolones as combination agents 2

When to Use Combination Therapy

• Septic shock or severe sepsis with Gram-negative bacteremia 2
• Suspected multidrug-resistant organisms based on local epidemiology 5
• Patients with prolonged hospitalization or recent antibiotic use 6
• Invasive material present (catheters, prosthetics) 3

Carbapenem-Resistant Organisms: Specialized Regimens

For Carbapenem-Resistant Enterobacterales (CRE)

• KPC-producing organisms: Meropenem-vaborbactam for bloodstream infections 7
• Metallo-β-lactamase producers: Ceftazidime-avibactam plus aztreonam plus metronidazole for intra-abdominal infections 7

For Carbapenem-Resistant Pseudomonas aeruginosa (CRPA)

• Severe pneumonia: Most active beta-lactam/beta-lactamase inhibitor combination in vitro, or cefiderocol combined with fosfomycin, plus aerosolized colistin or aminoglycosides 7
• Neurological infections: Ceftazidime-avibactam plus fosfomycin 7

For Carbapenem-Resistant Acinetobacter baumannii (CRAB)

• Bloodstream infections: Ampicillin-sulbactam plus colistin 7

Critical Pitfalls to Avoid

Do NOT Use These Agents

• Never use tigecycline for 3GCephRE infections (strong recommendation against) 1
• Avoid new beta-lactam/beta-lactamase inhibitors for 3GCephRE due to antibiotic stewardship—reserve these for extensively resistant bacteria 1
• Ceftazidime-avibactam has NO activity against Staphylococcus aureus—must add separate anti-staphylococcal coverage (vancomycin or linezolid) if MRSA is possible 8

Common Errors

• Using third-generation cephalosporins in areas with >18% resistance rates without susceptibility data 3
• Monotherapy in severe sepsis when combination therapy reduces inappropriate treatment from 36% to 22.2% 2
• Failing to adjust therapy based on local resistance epidemiology 5

De-escalation Strategy

Once patients are stabilized, step down from carbapenems to narrower agents based on susceptibility patterns. 1

• Switch to older beta-lactam/beta-lactamase inhibitors, quinolones, or cotrimoxazole when appropriate 1
• Individualized prediction models can facilitate early de-escalation without compromising adequacy (29% vs 21% de-escalation rate) 9
• Ensure adequate therapy time before de-escalation: median 4-5 hours to adequate therapy is acceptable 9

Dosing Optimization for Severe Infections

• Loading doses and continuous infusion of beta-lactams improve outcomes 5
• High-dose piperacillin-tazobactam (with loading dose and continuous infusion) for ESBL-producing Enterobacterales with MIC ≤8 mg/L 7
• Aminoglycosides (gentamicin 5-7 mg/kg/day IV once daily) should be administered after fluid resuscitation to reduce nephrotoxicity 4
• Adjust all antibiotic doses according to renal function 5

Local Epidemiology Matters

Selection of empiric regimens must be based on local resistance patterns, not just national guidelines. 1, 5

• Geographic variation is significant: Paris region vs. West France showed different resistance rates (58% vs 94% negative predictive value) 3
• Institutional sepsis protocols based on local epidemiology increase appropriate empirical treatment rates even when infectious disease consultation is available 5
• Scheduled rotation of antibiotic classes based on local resistance can reduce inadequate treatment from 6.1% to 4.5% 10