Back to the future: The story of colistin
In the early 1960s, colistin in the form of colistimethate and polymyxin B were used in this country to treat Pseudomonas aeruginosa infection because there were no other effective drugs. They were replaced by the anti-pseudomonal aminoglycosides, such as gentamicin, in the mid to late 1960s because of a perceived decrease in toxicity.
Now, with the emergence of P. aeruginosa, Acinetobacter baumannii and carbapenem-resistant Enterobacteriaceae (CRE), resistant to virtually all other antibiotics, colistimethate has again become an important agent.
With the increasing use of colistimethate, there have been pharmacokinetic and pharmacodynamic analyses in critically ill patients demonstrating that many of the old studies and the assumptions drawn from them were seriously flawed. As a result, the package inserts and the clinical use of colistimethate in many reports do not reflect what we feel to be the optimal use of this drug.
Keith S. Kaye
Colistimethate is a prodrug of colistin and, as such, has no intrinsic antibacterial activity. A vial of colistimethate contains 150 mg of what is called “colistin base activity (CBA),” which is equal to 5 million international units (MIU) of colistin. However, it actually has no activity until active colistin is released from the prodrug in vivo or in vitro. The package insert-based recommended dosing of colistimethate in the United States has been 2.5 mg/kg to 5 mg/kg CBA/day in two to four divided doses in patients with normal renal function, with a daily limit of 300 mg CBA.
Colistin slowly released
The rub is that, unlike most other antimicrobial prodrugs with which we deal, the release of active drug is very slow. Therefore, with a commonly used dose of 150 mg CBA IV, while blood levels of inactive colistimethate rapidly peak, colistin is only released slowly over time from the prodrug remaining in the serum; all of this time the prodrug is being excreted by the kidneys. The older studies used microbiological techniques to measure colistin activity and did not differentiate between active colistin actually present at the time of phlebotomy and the amount released from colistimethate during incubation in vitro.
More recent studies, such as those by Plachouras, published in 2009, and Garonzik, published in 2011, using liquid chromatographic methods after centrifugation of blood and processing at low temperatures have clarified the pharmacokinetics as follows.
The half-life in serum of colistimethate is 4 to 5 hours in critically ill patients with normal renal function. The half-life is prolonged in the presence of renal insufficiency, and active colistin continues to be released as long as colistimethate remains.
Active colistin has a half-life of about 9 hours and is not appreciably excreted in the urine. However, with renal insufficiency, the half-life is prolonged to about 13 hours because of the continuing release of colistin from the increased colistimethate still present in blood.
A therapeutic problem arises because the peak serum level of active colistin does not occur until about 7 hours after administration of colistimethate and is less than 1 mcg/mL. With standard maximum dosing of 150 mg CBA every 12 hours, the steady state peak serum level (level after 4 half-lives) of active colistin reaches a median level of 2 mcg/mL to 3 mcg/mL.
The pharmacodynamic data indicate that the AUC:MIC ratio appears to be the parameter best associated with bactericidal activity. The Clinical and Laboratory Standards Institute (CLSI) breakpoint for susceptibility of P. aeruginosa and A. baumannii is 2 mcg/mL or less, and heteroresistance is common in these organisms and in some Enterobacteriaceae. Furthermore, the protein binding by colistin is about 50%. Therefore, the initial peak of less than 1 mcg/mL (less than 0.5 mcg/mL unbound) is subtherapeutic for many strains of P. aeruginosa, and the steady state level of 2 mcg/mL to 3 mcg/mL (1 mcg/mL to 1.5 mcg/mL unbound) is borderline therapeutic. Furthermore, the steady state peak is not reached for a prolonged period of time after initiation of therapy.
Despite the package insert-recommended use of IV or IM colistimethate, because of the complex pharmacokinetics and lack of modern studies of IM administration, the drug should not be given by the IM route. In addition, because of the modern observations, investigators have made several recommendations for more appropriate dosing of colistimethate in critically ill patients. From review of their data, the following seems reasonable for use of colistimethate in the seriously ill patient.
First, an IV loading dose of 5 mg/kg CBA of ideal body weight not to exceed 300 mg CBA should be used. Second, animal work suggests that every 8-hour dosing is superior to every 12-hour regarding toxicity, and in vitro analyses suggest this more frequent administration might decrease the likelihood of development of resistance. With a creatinine clearance of 50 mL/minute or higher, 8 hours after the loading dose, we would recommend starting 5 mg/kg/day CBA using ideal body weight in three equally divided IV doses.
In the presence of a creatinine clearance of 30 mL/minute to 49 mL/minute, because of the prolonged half-life of colistimethate, after the loading dose, we would decrease the daily dose of colistimethate to 3.5 mg/kg/day CBA, divided into every 12 hour doses starting 12 hours after the loading dose. With a creatinine clearance of 10 mL/minute to 29 mL/minute, the dose would be 2.5 mg/kg/day CBA in doses every 12 hours, starting 12 hours after the loading dose. With creatinine clearance less than 10 mL/minute or in the setting of hemodialysis, the dose would be 1.5 mg/kg/day CBA, once daily starting 24 hours after the loading dose.
Higher doses equal better results
We realize that these doses of colistimethate are considerably higher than has commonly been used in the United Kingdom and in Europe, and somewhat higher than the doses recommended in the United States. However, these higher doses make sense when considering recently published pharmacokinetic and pharmacodynamic data. In fact, several studies have shown that, over a range of doses, better results have been obtained with the higher end of dosing.
Plachouras D. Antimicrob Agents Chemother. 2009;53:3430-3436.
For more information:
Keith S. Kaye, MD, MPH, Professor of internal medicine and infectious diseases in Wayne State University School of Medicine, and corporate director of Infection Prevention, Hospital Epidemiology and Antimicrobial Stewardship at the Detroit Medical Center. He is also an Infectious Disease News Editorial Board member.
Jason M. Pogue, PharmD, is a clinical pharmacist, Infectious Diseases, Sinai-Grace Hospital, Detroit Medical Center.
Disclosure: The columnists report no relevant financial disclosures.