LETTER
We read with interest the South African study based on a collection of separate case reports (1). The authors hypothesized that selective digestive decontamination (SDD) using enteral colistin (i.e., polymyxin E) and tobramycin probably contributed to the selection and carriage of a colistin-resistant OXA-181-producing Klebsiella pneumoniae strain. We argue that the use of an inappropriate protocol of enteral decontamination is more likely responsible for both the failure to eliminate intestinal carriage of K. pneumoniae and the selection of colistin-resistant K. pneumoniae.
Interestingly, the authors used their own decontaminating regimen that included a 2% colistin-tobramycin paste applied three times a day into the oropharynx, combined with a solution of 1 million IU colistin and 80 mg tobramycin administered three times a day into the gut via a nasogastric tube. Three main points should be addressed on this decontaminating protocol.
First, the dosage of the oropharyngeal paste is not specified. Generally, half a gram of 2% polymyxin E-tobramycin paste or gel four times a day has been shown to clear aerobic Gram-negative bacilli (AGNB) from the oropharynx (2).
Second, and most importantly, the enteral dose of 1 million IU colistin three times a day might be too low to achieve successful decontamination, i.e., eradication of the OXA-181-producing K. pneumoniae from the gut. Remarkably, the original SDD protocol uses 100 mg colistin sulfate four times a day (2). One milligram of colistin sulfate is equivalent to about 0.670 mg of pure colistin base (Fagron; colistin sulfate certificate of analysis, 2011), which has been assigned a potency of 30,000 IU per mg (3). Therefore, the original SDD protocol consists of about 2 million IU colistin four times a day. This daily dose is approximately three times higher than that used by the South African researchers. Early studies on SDD have shown that high doses of enteral colistin are required to suppress AGNB because colistin is moderately inactivated by fecal and food compounds (4, 5).
Third, although the investigators do not report it explicitly, the OXA-181-producing K. pneumoniae strain was almost certainly resistant to tobramycin, the second component of SDD. For successful decontamination, AGNB should ideally be sensitive to both decontaminating agents (2).
The failure to decontaminate the gut from AGNB may cause gut overgrowth, which is the ideal condition for the de novo development of new clones following increased spontaneous mutation, resulting in polyclonality (6). The low colistin levels in the gut will kill sensitive clones but allow mutating ones to become resistant to colistin.
Interestingly, in a recent large Dutch study of about 6,000 patients using the original protocol of SDD (7), the acquisition rates of colistin-resistant AGNB in the respiratory tract were low and comparable during SDD and standard care (8).
We believe that this study provides a basis for the conclusion that colistin resistance may be caused by the administration of an inappropriately low enteral dose of colistin. This is the decontamination regimen which should be strongly discouraged.
Footnotes
Ed. Note: The authors of the published article did not feel that a reply was necessary.
REFERENCES
- 1. Brink AJ, Coetzee J, Corcoran C, Clay CG, Hari-Makkan D, Jacobson RK, Richards GA, Feldman C, Nutt L, van Greune J, Deetlefs JD, Swart K, Devenish L, Poirel L, Nordman P. 2013. Emergence of OXA-48 and OXA-181 carbapenemases among Enterobacteriaceae in South Africa and evidence of in vivo selection of colistin resistance as a consequence of selective decontamination of the gastrointestinal tract. J. Clin. Microbiol. 51:369–372 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Silvestri L, de la Cal MA, van Saene HKF. 2012. Selective decontamination of the digestive tract: the mechanism of action is control of overgrowth. Intensive Care Med. 38:1738–1750 [DOI] [PubMed] [Google Scholar]
- 3. Falagas ME, Kasiakou SK. 2006. Use of international units when dosing colistin will help decrease confusion related to various formulations of the drug around the world. Antimicrob. Agents Chemother. 50:2274–2275 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. van Saene JJM, van Saene HKF, Tarko-Smit NJP, Beukeveld GJJ. 1988. Enterobacteriaceae suppression by three different oral doses of polymyxin E in human volunteers. Epidemiol. Infect. 100:407–417 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. van Saene JJM, van Saene HKF, Stoutenbeek CP, Lerk CF. 1985. Influences of faeces on the activity of antimicrobial agents for decontamination of the alimentary canal. Scand. J. Infect. Dis. 17:295–300 [PubMed] [Google Scholar]
- 6. van Saene HKF, Taylor N, Damjanovic V, Sarginson RE. 2008. Microbial gut overgrowth guarantees increased spontaneous mutation leading to polyclonality and antibiotic resistance in the critically ill. Curr. Drug Targets 9:419–421 [DOI] [PubMed] [Google Scholar]
- 7. de Smet AMGA, Kluytmans JAJW, Cooper BS, Mascini EM, Benus RFJ, van der Werf TS, van der Hoeven JG, Pickkers P, Bogaers-Hofman D, van der Meer NJM, Bernards AT, Kuijper EJ, Joore JCA, Leverstein-van Hall MA, Bindels AJGH, Jansz AR, Wesselink RMJ, de Jongh BM, Dennesen PJW, van Asselt GJ, te Velde LF, Frenay IHME, Kaasjager K, Bosch FH, van Iterson M, Thijsen SFT, Kluge GH, Pauw W, de Vries JW, Kaan JA, Arends JP, Aarts LPHJ, Sturm PDJ, Harinck HIJ, Voss A, Uijtendaal EV, Blok HEM, Thieme Groen ES, Pouw ME, Kalkman CJ, Bonten MJM. 2009. Decontamination of the digestive tract and oropharynx in ICU patients. N. Engl. J. Med. 360:20–31 [DOI] [PubMed] [Google Scholar]
- 8. Oostdijk EAN, Smits L, de Smet AMGA, Leverstein-van Hall MA, Kesecioglu J, Bonten MJM. 2012. Colistin resistance in gram-negative bacteria during prophylactic topical colistin use in intensive care units. Intensive Care Med. [Epub ahead of print.] doi:10.1007/s00134-012-2761-3 [DOI] [PubMed] [Google Scholar]