MDR Gram-Negative Risk Stratification Tools
While several risk scores exist to predict multidrug-resistant Gram-negative (MDR-GN) infections, no single validated tool has been universally adopted in clinical guidelines, and you should instead use a structured risk factor assessment combined with local epidemiology to guide empirical antibiotic selection.
Key Risk Factors for MDR-GN Infections
The most consistently identified risk factors across studies include 1, 2, 3, 4:
- Prior colonization or infection with MDR-GN (strongest predictor)
- Recent antibiotic exposure (especially carbapenems, broad-spectrum cephalosporins, fluoroquinolones within past 90 days)
- Prolonged hospitalization (≥7 days current admission or hospitalization within past 90 days)
- ICU admission (recent or current)
- Invasive devices (mechanical ventilation, central lines, urinary catheters)
- Severe immunosuppression (neutropenia, solid organ transplant, hematologic malignancy)
- Poor functional status
- Hemodialysis
- Severe mucositis (in cancer patients)
Available Risk Scores
For Cancer Patients
A 5-point clinical score was developed specifically for cancer patients with bacteremia 3:
- Recent antibiotic use (2 points)
- Recent ICU admission (2 points)
- Hospitalization ≥7 days prior to bacteremia (2 points)
- Severe mucositis (3 points)
- Prior MDR-GN colonization/infection (2 points)
Score ≥2 points: 66% sensitivity, 78% specificity for MDR-GN bacteremia. Score 0-1 points: negative likelihood ratio 0.18, post-test probability only 11.68% for MDR-GN.
General ICU Populations
Multiple scores exist (Giannella, Tumbarello, INCREMENT-CPE, CarbaSCORE) but have variable accuracy 4. These scores generally perform better at excluding low-risk patients than identifying high-risk ones.
Practical Clinical Approach
Use this algorithmic framework 1, 2, 5:
Step 1: Assess Local Epidemiology
- Know your institution's resistance rates for Enterobacterales, Pseudomonas aeruginosa, and Acinetobacter baumannii
- High-prevalence settings (>10-20% carbapenem resistance) warrant lower threshold for empirical coverage
Step 2: Identify High-Risk Patients
Consider empirical MDR-GN coverage if ≥2 major risk factors present:
- Prior MDR-GN colonization/infection (most important)
- Recent broad-spectrum antibiotics (within 90 days)
- Prolonged hospitalization (≥7 days or recent admission)
- ICU stay
- Severe immunosuppression
Step 3: Severity Assessment
Empirical MDR-GN coverage is more justified when:
- Septic shock present
- Severe neutropenia (<100 cells/μL)
- Solid organ transplant recipient
- Delayed appropriate therapy would be catastrophic
Step 4: Site-Specific Considerations
- Respiratory tract: Higher risk with P. aeruginosa and Acinetobacter in ventilated patients
- Urinary tract: Consider colonization patterns; may allow narrower empirical therapy if no sepsis
- Bloodstream: Highest stakes; appropriate empirical therapy critical for mortality reduction 6
Critical Caveats
- Scores predicting colonization have low accuracy
- Scores predicting infection among colonized perform better but still imperfect
- Most scores were developed in specific populations (cancer, ICU) and may not generalize
- Colonization does not equal infection 7—clinical context determines relevance
Common pitfall: Treating colonization rather than infection. Positive surveillance cultures should inform empirical choices when infection is suspected, not trigger automatic treatment 7.
Guideline-Supported Empirical Regimens
When MDR-GN risk is high, guidelines support 8, 6, 9:
For suspected MDR-GN with sepsis/severe infection:
- Antipseudomonal β-lactam (piperacillin-tazobactam, cefepime, ceftazidime, or carbapenem) PLUS
- Second agent (aminoglycoside or fluoroquinolone) for Pseudomonas coverage
- Add vancomycin or linezolid if MRSA risk factors present
For carbapenem-resistant organisms (when specifically suspected):
- Reserve newer agents (ceftazidime-avibactam, meropenem-vaborbactam, cefiderocol) for targeted therapy based on culture results 9, 10, 5
- Empirical use only justified in highest-risk scenarios with documented prior CRE colonization and septic shock
The key principle: Risk stratification should guide whether to cover MDR-GN empirically, but rapid de-escalation based on culture results and clinical response is essential to preserve these agents 1, 9.