Colistin Dosing and Management for Multidrug-Resistant Gram-Negative Infections
For adults with confirmed carbapenem-resistant gram-negative infections when no safer alternatives exist, administer a loading dose of 9 million IU (approximately 5 mg/kg) of colistimethate sodium (CMS), followed by a maintenance dose of 4.5 million IU every 12 hours in patients with creatinine clearance >50 mL/min. 1
Standard Dosing Algorithm
Loading Dose (Critical for All Patients)
- 9 million IU CMS (equivalent to ~300 mg colistin base activity)
- Administer regardless of renal function 2
- Required because CMS converts slowly to active colistin, taking 2-3 days to reach steady state without loading 2
Maintenance Dosing Based on Renal Function
For CrCl >50 mL/min:
For CrCl 30-49 mL/min:
- 2.5 mg/kg once daily or divided into 2 doses 3
For CrCl 10-29 mL/min:
- 1.5 mg/kg every 36 hours 3
For continuous renal replacement therapy (CRRT):
For intermittent hemodialysis:
- 2 million IU every 12 hours after normal loading dose 2
- Schedule dialysis toward end of dosing interval
Administration Methods
The FDA label provides two acceptable IV approaches 3:
- Direct intermittent: Inject half the total daily dose over 3-5 minutes every 12 hours
- Continuous infusion: Inject first half over 3-5 minutes, then infuse remaining half over 22-23 hours
Extended infusion may optimize pharmacodynamics, particularly for organisms with higher MICs 2.
Combination Therapy Considerations
Combination therapy remains controversial with weak evidence 1. The 2022 guidelines note that colistin-meropenem combination was not superior to monotherapy in one RCT (14-day mortality 25% vs 31%, p=1.0) 1. However:
- Consider combination when pathogen shows in vitro susceptibility to additional agents 1
- If no susceptible second agent exists, combine with carbapenem at lowest MIC 1
- For CRAB bloodstream infections specifically, colistin-carbapenem combination is weakly recommended 1
- The 2023 Chinese guidelines suggest polymyxin-carbapenem combinations when meropenem MIC ≤8 mg/L with high-dose extended infusion 4
The European 2022 guidelines provide high-certainty evidence AGAINST carbapenem-polymyxin combinations for CRAB when isolates have typical high-level resistance (MIC >16 mg/L) 5.
Mandatory Monitoring
Renal function must be monitored closely throughout therapy 1. This is a strong recommendation despite low-quality evidence because:
- Acute kidney injury is a major factor in clinical failure and mortality 1
- Nephrotoxicity occurs in 15-33% of patients 2
- Risk is higher in elderly and those with pre-existing chronic kidney disease 1
Monitor creatinine clearance at minimum every 2-3 days, adjusting maintenance doses accordingly.
Critical Pitfalls to Avoid
Omitting the loading dose: Without it, therapeutic concentrations take 2-3 days to achieve, during which mortality risk increases 2
Underdosing patients with normal/augmented renal clearance: Standard regimens may provide inadequate exposure when MIC ≥1 mg/L in patients with preserved renal function 6, 7
Using colistin monotherapy for high-MIC organisms: When MIC >1 mg/L, combination therapy should be strongly considered as dose escalation risks nephrotoxicity 8
Failing to adjust for obesity: Dose on ideal body weight, not actual weight 3
Continuing therapy without clinical response: Mortality remains 36-90% across studies despite treatment 9, 10; reassess source control and consider alternative/additional agents if no improvement by 3-5 days
Therapeutic Drug Monitoring
Where available, TDM is warranted due to high interpatient variability 8, 11. Target steady-state plasma concentrations of 2-2.5 mg/L for systemic infections, though achieving pharmacodynamic targets does not guarantee clinical success in severely ill patients 7.
Special Populations
For adjunctive inhaled colistin in pneumonia: 1.25-15 million IU daily divided every 8-12 hours, each dose diluted in 5 mL sterile saline 12. The 2023 guidelines weakly recommend adding aerosolized polymyxin to IV therapy for respiratory tract infections when clinical response is suboptimal 4.