Global survey of polymyxin use: A call for international guidelines.

Polymyxins (polymyxin B and colistin) are older bactericidal antibiotics that are increasingly used to treat infections caused by multidrug-resistant (MDR) Gram-negative bacteria. However, dosing and clinical use of these drugs vary widely. This survey was undertaken to reveal how polymyxins are used worldwide. Data were collected through a structured online questionnaire consisting of 24 questions regarding colistin usage patterns and indications as well as colistin dosage for adult patients. The questionnaire was disseminated in 2011 to relevant experts worldwide and was completed by 284 respondents from 56 different countries. Respondents from 11/56 countries (20%) had no access to colistin; 58/284 respondents (20.4%) reported that in 2010 they experienced that colistin was not available when needed. Formulations of polymyxins used were reported as: colistimethate sodium (48.6%); colistin sulfate (14.1%); both (1.4%); polymyxin B (1.4%); and unknown. Intravenous formulations were used by 84.2%, aerosolised or nebulised colistin by 44.4% and oral colistin for selective gut decontamination by 12.7%. Common indications for intravenous colistin were ventilator-associated pneumonia, sepsis and catheter-related infections with MDR Gram-negative bacteria. Only 21.2% of respondents used a colistin-loading dose, mainly in Europe and North America. This survey reveals that the majority of respondents use colistin and a few use polymyxin B. The survey results show that colistin is commonly underdosed. Clear guidance is needed on indications, dosing and antibiotic combinations to improve clinical outcomes and delay the emergence of resistance. Colistin should be considered a last-resort drug and its use should be controlled. International guidelines are urgently needed.

Introduction

Severe multidrug-resistant (MDR) Gram-negative infections are increasing worldwide [1]. The emergence of carbapenem resistance in Gram-negative bacteria is of extreme concern as few therapeutic options remain [1]. For this reason, clinicians are increasingly using an older class of antibiotics, namely the polymyxins, most commonly colistin (polymyxin E) [2].

Colistin is a bactericidal antibiotic with a broad Gram-negative spectrum, consisting of a mixture of colistin A and colistin B, differing in their fatty acid side chain. The active compound was isolated from the bacterium Bacillus polymyxa var. colistinus in 1949 and was used in patients for the first time in 1959. Polymyxins have been used rarely since the 1970s when less toxic aminoglycosides and other antibiotics came on the market [3]. Polymyxins, developed over 50 years ago, have not been subjected to the rigorous studies to optimise dosing regimens and to demonstrate efficacy that are currently required by regulatory agencies for new drugs [3]. The appropriate dosing schedule for these drugs, how long they should be administered, and with which other antibiotics they should be used are questions remaining to be answered [2-5]. Insights into pharmacodynamic properties and dose–effect relationships have only recently be partly elucidated, but many questions still remain [6]. The lack of documented experience with the drug and the availability of various forms of polymyxins with different concentrations and even units used in different dosing schemes have also led to inappropriate use. Furthermore, access to polymyxin drugs and other ‘forgotten antibiotics’ is limited – even in developed nations – for economic reasons [7].

Considering that no new antibiotics covering MDR Gram-negative bacteria can be expected to reach the market in the near future, it is essential to use the polymyxin class of antibiotics optimally and rationally [8,9]. Providing early adequate therapy is critical in patients with severe infections caused by MDR bacteria [8,9]. Underdosing runs the risk of treatment failure, poor outcome and potentially the development and spread of polymyxin resistance [6]. It is currently unknown which doses clinicians use and whether they are using colistin loading doses. It is also not known with which other drugs colistin is being combined.

To assess global polymyxin availability and parenteral use practices, an online survey was conducted.

Methods

To assess global systemic polymyxin use, a 24-question survey (see Appendix) was developed that sought information on: characteristics of the respondents; indications for use of polymyxins; access to polymyxins and drug type; cost of polymyxins; dosing of polymyxins; adverse events; antibiotic combinations; and research needs. Topical polymyxin use was not covered in this survey.

The questionnaire was peer-reviewed by three working groups working on antibiotic resistance: (i) Global Antibiotic Resistance Partnership (GARP); (ii) Action on Antibiotic Resistance (ReAct); and (iii) the Antimicrobial Stewardship Working Group of the International Society of Chemotherapy. The final version was piloted and then made freely available through an online survey website (http://www.freeonlinesurveys.com).

Professionals (e.g. infectious diseases doctors, pharmacists, microbiologists, intensive care physicians) in relevant networks were invited by email to complete the questionnaire. All invitees were asked to forward the email to anyone they considered relevant to complete the survey. The survey was open from 1 June to 1 November 2011. Data were analysed using descriptive statistics using SPSS v.15 (SPSS Inc., Chicago, IL).

Results

Respondent characteristics

The survey was completed by 284 respondents from 56 different countries (Table 1 and Fig. 1). The majority of respondents were from Europe, followed by the Americas, and most worked in tertiary care teaching hospitals. Approximately one-half of the respondents worked in hospitals with more than 500 beds. Most of the hospitals had an intensive care unit (89.4%), department of surgery (92.6%) and a microbiology laboratory (91.9%). All respondents had a medical background relevant to the survey (Table 1).

Table 1

Characteristics of the respondents regarding country, hospital and profession (N = 284).

Characteristic	n (%)
World regions
Europe	115 (40.5)
South America	64 (22.5)
North America	33 (11.6)
Oceania	35 (12.3)
Asia	26 (9.2)
Africa	11 (3.9)
Hospital typea
Primary care	47 (16.5)
Secondary care	74 (26.1)
Tertiary care	205 (72.2)
Teaching	185 (65.1)
Profession
Infectious diseases doctor	144 (50.7)
Clinical microbiologist	35 (12.3)
Intensive care doctor	50 (17.6)
Pharmacist	22 (7.7)
Other	33 (11.6)

Respondents were able to select more than one possibility, therefore figures do not add to 100% (e.g. tertiary hospital and teaching are not mutually exclusive; some respondents worked at several places).

Fig. 1

World map showing the locations of the respondents. The map includes information on the availability of colistin and whether the respondent uses a colistin-loading dose (more detailed information by respondent available from the corresponding author by request).

Polymyxin drug access

Respondents in 11 (20%) of 56 countries reported that they had no access to colistin at the time of the survey. Lack of access was reported from all continents except Oceania and North America. The countries for which no access was reported, included Bolivia, Guatemala, Indonesia, Laos, Norway, Portugal, Russia, Uzbekistan, Venezuela, Vietnam and Yemen (Fig. 1; data by respondent available from the corresponding author by request). The survey also asked about the consistency of supply of polymyxin drugs in the respondents’ institutions. Of the 284 respondents, 58 (20.4%) reported that these drugs were unavailable for patients at least once during the preceding year (Fig. 1).

Various forms of polymyxin drugs are available and used worldwide. The majority of respondents used colistimethate sodium (48.6%), followed by colistin sulfate (14.1%), both forms of colistin (1.4%) and polymyxin B (1.4%), and the remainder did not know the exact formulation. Eighty percent of the reported colistin sulfate use originated from Europe and South America. No colistin sulfate was reported from Asia. Polymyxin B was used only in South America (Brazil and Panama) and Asia (Singapore). As Polymyxin B is rarely used, further analysis here is limited to colistin.

The number of patients [median; interquartile range (IQR)] treated with polymyxin drugs by region in the year 2010 was: 30 (IQR 100) in Africa; 32 (IQR 300) in Asia; 5 (IQR 24) in Europe; 2 (IQR 20) in North America; 2 (IQR 4) in Oceania; and 15 (IQR 30) in South America.

Indications for colistin

Colistin is administered via different routes, depending on the indication. Of the 284 respondents, 84.2% used colistin intravenously, 44.4% used aerosolised or nebulised colistin and 12.7% used oral colistin for selective gut decontamination. Common indications for intravenous colistin use were ventilator-associated pneumonia, sepsis and catheter-related infections with MDR Gram-negative bacteria. Nebulised colistin was often used in cystic fibrosis patients. Colistin was used most commonly against Acinetobacter baumannii and Pseudomonas aeruginosa. Oral colistin for selective gut decontamination was most frequently reported from Europe, followed by North America. No African respondents reported oral colistin use.

Dosing of colistin

The dosing schedule for an adult patient with a weight of 70 kg and normal renal function (the standard case posed in the questionnaire) varied considerably, partly reflecting the different types of formulations. Dosing data were reported in million international units (MIU) (37.3%), mg/kg/day (21.5%) and mg (8.8%) (see Table 2 for dosing conversion). For those reporting in MIU, the median total daily dose was 6.0 MIU (range 1.5–9.0 MIU) administered three times per day (range 1–4 times per day). For those reporting in mg/kg/day, the median colistin dose was 5 mg/kg/day (range 1.5–6.0 mg/kg/day). Those who also administered a loading dose (see below) reported a higher maintenance dose compared with those who did not use a loading dose: median of 9 MIU versus 6 MIU. Approximately one-quarter (26.4%) of the respondents were dosing their patients in the lower range (≤2.5 mg/kg/day or ≤6 MIU). The median duration of colistin treatment was 14 days. Less than one-third (29%) reported adjusting the dose in obese patients. Dosing strategies for patients on haemodialysis varied, most indicating that they halved the dose.

Table 2

Conversion table milligrams to international units (IU) for someone weighing 70 kg.

1 mg colistin base activity ≅33,250 IU

5 mg/kg colistin base activity = 350 mg/day ≅11.5 MIU/day

1 mg colistimethate sodium ≅12,500 IU

9 MIU ≅720 mg colistimethate sodium ≅3.9 mg/kg colistin base activity

MIU, million international units.

Evidence from pharmacokinetic/pharmacodynamic (PK/PD) studies indicates that a colistin loading dose may be beneficial for patients with severe MDR Gram-negative infections. Of the respondents, only 21.2% reported a loading dose and most of these were in Europe and North America. No Asian respondent reported that they used a loading dose. In Europe, the most common loading dose was 9 MIU and in South America it was 4.5 MIU (the loading dosage was provided via a separate email request to those using a loading use).

Drug combinations

Table 3 lists the antibiotics used with colistin, according to the relative frequency. The most common antibiotic combination was with a carbapenem. Other relatively common combinations were aminoglycosides, tigecycline, rifampicin and piperacillin/tazobactam. According to respondents, the choice of a second antibiotic depends on the type of infection being treated and the susceptibility profile of the cultured bacteria, if available. The most important drugs to combine with polymyxins to study in a clinical trial are considered to be tigecycline and piperacillin/tazobactam, according to a majority of survey respondents.

Table 3

Response to what drugs are used for combination therapy with colistin.

	Frequency (%)
	Never	Uncommon	Common
Carbapenem	24.5	31.7	43.8
Gentamicin	41.3	38.8	20.0
Tigecycline	47.4	34.2	18.4
Rifampicin	51.9	29.8	18.3
Piperacillin/tazobactam	49.3	35.5	15.2
Fluoroquinolone	52.8	39.6	7.7
Fosfomycin	81.5	12.0	6.5
Minocycline	81.1	14.5	4.4
Othera	64.3	23.1	12.6

Other includes cephalosporins, trimethoprim/sulfamethoxazole, glycopeptide, linezolid and doxycycline.

Discussion

To our knowledge, this is the first global survey on polymyxin use. Respondents were a self-selected group and we do not portray the results as an unbiased sample. Despite this, the study provides important insights into the current global pattern of access to and use of these drugs. As only a few respondents reported polymyxin B use, the discussion focuses on colistin.

In general, the indications for use of systemic colistin are similar across the regions: severe infections caused by MDR Gram-negative bacteria, such as A. baumannii and P. aeruginosa. However, there are major variations in the dosing regimens, with several sites using relatively low doses, and a loading dose of colistin was not reported by most respondents. PK/PD studies found that it takes several days to develop adequate concentrations of colistin in human plasma [5,10]. These studies suggest that loading doses are required to achieve adequate levels of colistin as early as possible [5,10]. Providing early adequate therapy is critical in patients with severe infections due to MDR bacteria [10].

In a recent discussion on dosing of colistin, and taking into account the PK/PD properties of colistin, a 9 MIU loading dose and 4.5 M IU every 12 h was suggested to be one of the optimum dosing alternatives (colistin fact sheet available at http://aida-project.eu). In particular, the loading dose was regarded to be essential, since the simulations strongly indicate that it takes up to several days before steady state is reached [10]. However, even then concentrations required for optimal efficacy may not be reached. In addition, PK profiles vary widely between patients, suggesting that therapeutic drug monitoring could be beneficial.

Clinicians reported reluctance to use higher doses because of concerns about toxicity [3]. However, based on various studies, these concerns may be misplaced and respondents that use a loading dose confirm little toxicity issues of this strategy. Colistin has been shown to have a better safety profile than thought previously, perhaps even better than the safety profile of aminoglycosides [3,11,12]. The risk of toxicity may be favoured above inadequate treatment of a severe infection due to MDR bacteria. Clearly, more should be known about these drugs, but we are concerned that underdosing may be a serious issue as this is related to the emergence of resistant bacteria [13].

However, excessive use of colistin needs to be addressed. More than 10% of the respondents mentioned that they also used polymyxins as part of a selective gut decontamination regimen. The use of polymyxins for this purpose needs to be weighted by the risk of resistance development in a world where carbapenem resistance is emerging and spreading [1]. Furthermore, colistin is used in large amounts in agriculture, particularly in Asia (Do Thuy Nga, GARP co-ordinator Vietnam, personal communication). The reasons for use in agriculture require further investigation and a case made for banning polymyxins for agricultural use.

This survey revealed important variations in the use of polymyxins across the world, supporting the need for the development of clear guidelines covering the indication, dosage and duration of polymyxins. Many questions regarding dosing and combinations with other antibiotics remain, and studies are under way to answer at least some of these. However, guidelines can and should be developed based on current evidence and revised as new information becomes available. The European Union-funded AIDA project tries to address these issues and has made colistin fact sheets and dosing conversion tables available (http://aida-project.eu/back-ground-information/fact-sheets/91-fact-sheets-public/84-colistin-fact-sheet-public). As new classes of antibiotics for MDR Gram-negative bacteria are not within sight, we need to preserve these last-resort antibiotics.

Funding

This work was funded by the Global Antibiotic Resistance Partnership (GARP, USA) and the Wellcome Trust (UK). The Wellcome Trust had no role in the design, data collection, analysis, interpretation, drafting of the paper or the decision to submit for publication. GARP staff were part of the study team and therefore had a role in all aspects of this project.

Competing interests

None declared.

Ethical approval

Not required.