Colistin, mechanisms and prevalence of resistance

Abstract

Background:

Infections caused by multi-drug-resistant Gram-negative bacteria, particularly Acinetobacter baumannii, Pseudomonas aeruginosa and Klebsiella pneumoniae, that cause nosocomial infections, represent a growing problem worldwide. The rapid increase in the prevalence of Gram-negative pathogens that are resistant to fluoroquinolones and aminoglycosides as well as all β-lactams, including carbapenems, monobactam, cephalosporins and broad-spectrum penicillins, has prompted the reconsideration of colistin as a valid therapeutic option. Colistin is an old class of cationic, which act by disrupting the bacterial membranes resulting in cellular death. Although there has been a significant recent increase in the data gathered on colistin, focusing on its chemistry, antibacterial activity, mechanism of action and resistance, pharmacokinetics, pharmacodynamics and new clinical application, the prevalence of colistin resistance has been very little reported in the literature. This review concentrates on recent literature aimed at optimizing the clinical use of this important antibiotic.

Methods:

The available evidence from various studies (microbiological and clinical studies, retrieved from the PubMed, and Scopus databases) regarding the mechanisms and prevalence of resistance was evaluated.

Results:

Increasing use of colistin for treatment of infections caused by these bacteria has led to the emergence of colistin resistance in several countries worldwide. Although resistance to polymyxins is generally less than 10%, it is higher in the Mediterranean and South-East Asia (Korea and Singapore), where colistin resistance rates are continually increasing.

Conclusion:

There is a critical need for effective infection prevention and control measures and strict use of antibiotics in the world to control the rise and spread of colistin resistance.

Keywords: :

• Colistimethate sodium
• Colistin
• Colistin resistance
• MDR Gram-negative bacteria
• Polymyxin
• Prevalence

Introduction

The world is facing an enormous and growing threat from the emergence of Gram-negative bacteria such as Pseudomonas aeruginosa, Acinetobacter baumannii, and Enterobacteriaceae. Development of resistance to most available antibiotics including β-lactams, carbapenems, fluoroquinolones, and aminoglycosides is increasing worldwide at an alarming rate. Therefore the lack of development of new antimicrobial agents has prompted the medical community to re-evaluate the use of colistin in many health care centers around the world, where no other less toxic or effective antibiotic is available1–3. In the past years, interest has been rekindled in the ‘old’ polymyxin antibiotic, colistin (polymyxin E), for the treatment of infections caused by multi-drug-resistant Gram-negative bacteria (MDR-GNB) owing to its favorable properties of rapid bacterial killing, a narrow spectrum of activity, and an associated slow development of resistance4. However, these salvage antibiotics were not available in the market in some countries for the treatment of patients with infections due to (inappropriately called) pan-drug-resistant Gram-negative bacteria, since isolates were susceptible to polymyxins5.

Most studies report that resistance to carbapenems is increasing and, since tigecycline has not been registered in many countries, colistin is often the only effective antibiotic against multi-drug-resistant organisms4. However, it is most notable for colistin activity toward MDR organisms including A. baumanni, P.aeruginosa and K. pneumonia, and current low levels of resistance have undergone as revival agents for treatment of infections caused by Gram-negative organisms.1^,6^,7.

As described in the ‘Bad Bugs, No Drugs’ paper published by the Infectious Diseases Society of America (IDSA), “as antibiotic discovery stagnates, a public health crisis brews”8. Therefore, there is an urgent need for new antibiotics, particularly those active against Gram-negative ‘superbugs’. This study research was done in order to optimizing the clinical use of this important antibiotic. To achieve this goal, almost all articles about microbiological and clinical studies in PubMed and Scopus databases regarding the mechanisms and prevalence of resistance were evaluated.

Early clinical experiences of colistin

One of the few remaining options left for treatment of life-threatening infections caused by MDR bacteria is colistin (or polymyxin E). It was first isolated in Japan by Koyama and co-workers from the spore-forming soil bacterium Bacillus polymyxa subsp. colistinus in 1947, and first used as an intravenous formulation in the 1950s9–11. Colistin has been approved by the US FDA and has been available since 1959 for the treatment of infections caused by Gram-negative bacteria, because these agents are rapidly bactericidal against them12^,13. During the ensuing decades, colistin was used in the treatment of several types of infections, including infectious diarrhea and urinary tract infections, as well as in bowel decontamination. Furthermore, polymyxins have been used for several decades in topical formulations for eye and ear infections and in regimens for selective bowel decontamination14. However, the high rates of toxicity associated with the use of polymyxins prompted their replacement by newer antibiotics like gentamicin and carbenicillin, two safer and more effective agents. Since the mid-1990s, there has been a greatly renewed interest in their clinical use due to prevalent MDR Gram-negative bacteria (especially P. aeruginosa, A. baumannii and K. pneumoniae) and lack of novel antibiotics8. Recently, they were relegated to the status of a reserve agent after early reports of a ‘high’ incidence of toxicity. When the use of a β-lactam, aminoglycoside, or quinolone is ineffective, the polymyxins, particularly colistin, remain drugs of last resort for the treatment of infections with MDR Gram-negative superbugs11.

Diversity and chemistry of colistin

Colistin, a polypeptide antibiotic of the polymyxin family, non-ribosomally synthesized with 1750 Da molecular weight, consisting of a cyclic heptapeptide with a tripeptide side chain acylated at the N terminus by a fatty acid tail (Figure 1). It is noteworthy that the hydrophobicity of the N-terminal fatty acyl segment is responsible for the inherent toxicity and greatly influences the antimicrobial activity of colistin15.

Figure 1. Colistin structures. (A) Structures of colistin A and B. (B) Structures of colistimethate A and B. Fatty acid: 6-methyloctanoic acid for colistin A and 6-methylheptanoic acid for colistin B; Thr = threonine; Leu = leucine; Dab = α,γ-diaminobutyric acid; α and γ indicate the respective amino groups involved in the peptide linkage. (Modified from Li et al.4).

Polymyxin includes five different chemical compounds (polymyxins A, B, C, D, and E). It is mainly two of them, colistin A (polymyxin E1) and colistin B (polymyxin E2), that have been used in clinical practice; these differ only in their fatty acid tails1^,16. The difference between polymyxin B and colistin lies in the amino-acid components17 and both are polycations at physiological pH owing to the five L-α,γ-diaminobutyric acid (Dab) residues18. There are two different forms of colistin available commercially, colistin sulfate for oral and topical use and colistimethate sodium (CMS) for parenteral and aerosol therapy; both forms may be given by inhalation. Colistin (usually used as the sulfate salt) is a polycation, whereas colistimethate (used as the sodium salt) is a polyanion at physiological pH19.

CMS is an inactive prodrug of colistin and has no intrinsic antibacterial activity. CMS produced by a sulfomethylation reaction in which the primary amine groups on α,γ-diaminobutyric acid (Dab) are reacted with formaldehyde followed by sodium bisulfite16^,20. The composition must be converted to colistin in vivo, but this occurs slowly and incompletely3. CMS is increasingly the last line of defense for multi-drug-resistant Gram-negative bacteria and is now being used as ‘salvage’ therapy in critically ill patients21. CMS is less toxic than colistin when it is administered parenterally22^,23. As opposed to CMS which is rapidly hydrolyzed in plasma, colistin is stable for several days at room temperature24. There are several commercial preparations of colistimethate, and their differences have undoubtedly contributed to confusion when evaluating dosing guidelines. ‘Coly-Mycin M’ Parenteral is produced by Parkedale Pharmaceuticals in the United States. Another preparation of colistimethate is ‘Colomycin Injection’, manufactured by Alpharma ApS (Denmark)25. Topical formulations of colistin sulfate and polymyxin B sulfate are also available in many countries. CMS is much more commonly used internationally (e.g. North America, South America, Asia, Europe and Australia) whereas parenteral polymyxin B is mainly available in the USA, Brazil and Singapore18. Nowadays, novel derivatives of polymyxins have been developed. Some of them, including NAB739, are directly antibacterial whereas others, including NAB7061, lack the direct activity but sensitize bacteria to other antibiotics26.

Pharmacokinetics/pharmacodynamics

There is a dearth of reliable information concerning the pharmacokinetic (PK) and pharmacodynamic (PD) data available to guide dosing in humans especially critically ill patients, including those on renal replacement therapy27. Most knowledge on the pharmacokinetics of colistin was obtained at least two decades ago when non-specific microbiological assays were used to measure the concentrations of colistin in biological fluids11. A better understanding of the PK/PD relationship of colistin is urgently needed to determine the optimal dosing regimen. Specifically, colistin concentrations in plasma of patients with preserved renal function have often been reported to be lower than expected and would be unlikely to be effective but may encourage resistance24.

The differences in chemistry between colistimethate and colistin also translate into differences in PK and PD19. Colistin is a concentration-dependent antibiotic and if in the fermentation process the production of the subcomponent peptides varies in amount and composition, it could conceivably have an impact on antibacterial activity or on the synergistic effects that are apparent with complex antibiotics that contain mixed active antibiotic substances (Figure 2)15.

Figure 2. Pharmacokinetic (PK) simulation of colistin A and B in the in vitro PK model study (mean ± SD, n = 3). (From Zhao et al.151).

Colistin sulfate and colistimethate sodium are poorly absorbed from the adult gastrointestinal tract, mucosal surfaces, inflamed surfaces or burns11. Therefore to achieve therapeutic serum level, intramuscular administration of colistimethate sodium must be employed. However, intravenous (i.v.) administration is reported to have a relatively poor Cerebrospinal fluid (CSF) distribution and clinical outcomes vary28. Under different conditions of temperature and time, different proportions of colistimethate sodium are hydrolyzed to colistin. After administration of CMS, colistin appears in plasma rapidly, approximately 50% bound to human plasma. Levels appeared higher but declined more rapidly than those achieved after intramuscular (i.m.) administration29.

A study performed to assess steady state serum concentrations of colistin after i.v. administration of colistin methanesulfonate (CMS) in critically ill patients revealed CMS dosage regimens administered to them were associated with suboptimal Cmax/minimum inhibitory concentration (MIC) ratios for many strains of Gram-negative bacilli currently reported as sensitive (MIC ≤ 2 mg/L)30. Evidence from PK/PD studies indicates that a colistin loading dose may be beneficial for patients with severe MDR Gram-negative infections31. However, the optimal dosing regimen of CMS was not established according to today’s rigorous drug development procedure, for CMS a two-compartment model and for colistin a one-compartment model best described the pharmacokinetics. The half-lives of the two phases for CMS were estimated to be 0.026 and 2.2 h, respectively and for colistin the estimated half-life was 18.5 h32; however, some data estimated it with slight differences (i.e. 14.4 h)33. A reported finding was a colistin plasma peak concentration of 2.93 mg/mL, occurring on average 15 min after the end of the 30 min infusion. However, such an early peak was not observed in the most recent studies, and it can most likely be explained, at least in part, by uncontrolled post-sampling CMS degradation, as only a low percentage of CMS hydrolysis would have a pronounced influence on early colistin concentrations34. The mean terminal half-life of colistin was 55.7 min. Approximately 60% of the dose was eliminated via the urine in 24 h and presented as a mixture of CMS and colistin21. This was also demonstrated by the rapid initial killing (<2 h) in the time–kill studies4. In contrast in a previous report on the in vitro pharmacodynamics of colistin against multi-drug-resistant P. aeruginosa, there was significant regrowth at 24 h35. Recent data from PK/PD population studies have suggested that this population could benefit from administration of higher than standard doses of CMS, but the relationship between administration of incremental doses of CMS and corresponding PK/PD parameters as well as its efficacy and toxicity have not yet been investigated in a clinical setting36.

During the first 24 h after dosing, colistin is exerted in the urine, and excessively high serum concentration may therefore occur in patients with renal insufficiency37. These patients might be expected to metabolize the drug more slowly than those with normal renal function. Studies show that in patients with severe renal failure who are not being dialyzed significant levels of colistin may still be present in the blood two to three days after a single dose of 2–3 mg/kg body weight intravenously38. Although PK/PD data in intensive care unit (ICU) patients are scarce, recent evidence shows that the PK and PD of CMS and colistin in critically ill patients differ from those previously found in other groups, such as cystic fibrosis (CF) patients39.

Spectrum of activity

The antimicrobial spectrum of colistin is narrow; it has in vitro activity against some MDR Gram-negative pathogens, including common or important non-fermentative Gram-negative bacteria, A.baumannii and P. aeruginosa11. The polymyxins have bactericidal activity against most members of the Enterobacteriaceae family includes E. coli, Enterobacter, Salmonella, Shigella and Klebsiella40. According to the European Committee on Antimicrobial Susceptibility Testing, P. aeruginosa and A. baumannii susceptibilities are defined as MICs of ≤4 and ≤2 mg/L colistin sulfate, respectively41, and according to the Clinical and Laboratory Standards Institute (CLSI) as an MIC of ≤2 mg/L for both bacteria37.

On the other hand eukaryotic microbes and mammalian cells, Proteus species, Neisseria species, Serratia species, Providencia species, Burkholderia pseudomallei, Morganella morganii, and Edwardsiella tarda as well as anaerobic bacteria are typically resistant to colistin42–47. The polymyxins also demonstrated no activity against Gram-negative and Gram-positive cocci and Gram-positive bacilli40. Stenotrophomonas maltophilia and Aeromonas species (except A. jandaei) are susceptible, although A. hydrophila has inducible resistance48^,49.

Mechanism of action

The bactericidal effect of colistin is extremely rapid but the exact mechanism by which colistin can kill bacterial cells is currently unclear. Between colistin and polymyxin B, colistin is used more extensively worldwide due to its better availability50. Colistin resistance is not dependent upon bacterial metabolic activity and acquired resistance is rare48. Gram-negative bacteria are characterized by the presence of an outer membrane. The protective function of the outer membrane mainly relies on the presence of lipopolysaccharide (LPS) constituents at the surface of the cell, which limit the penetration of hydrophobic and/or large antibiotics51. The polyanionic bacterial LPS is the initial target, it bears negative charge and this LPS confers to the integrity and stability of the bacterial outer membrane. But polymyxins having positive charge is critical for their interaction with the hydrophobic lipid A component of LPS52^,53. Their antibacterial effect on Gram-negative bacteria acts through a detergent-like effect, via a two-step mechanism. It comprises initial binding with electrostatic interactions between the polycationic ring of colistin to cell envelope components, causing the displacement, in a competitive fashion, of the calcium (Ca²⁺⁾ and magnesium (Mg²⁺) ions from the phosphate groups of LPS that act as membrane stabilizers, leading to disruption of the outer membrane and to the loss of cellular contents, thus killing the bacterium (Figure 3)54–56. This process is independent of the entry of polymyxins into the cell57, and seems to be inhibited in the presence of these bivalent cations53. The killing process with colistin is not dependent upon bacterial metabolic activity, and this may be a significant contributing factor towards the slow development of resistance, a resistance which develops more slowly than that to tobramycin11.

Figure 3. Action of colistin on bacterial membrane. The cationic cyclic decapeptide structure of colistin binds with the anionic LPS molecules by displacing calcium and magnesium from the outer cell membrane of Gram-negative bacteria, leading to permeability changes in the cell envelope and leakage of cell contents. By binding to the lipid A portion of LPS, colistin also has an anti-endotoxin activity. Disruption of the membranes should promote permeability for more conventional anti-pseudomonals. LPS: lipopolysaccharides; PBP: penicillin-binding protein. (From Martis et al.58).

Besides leading to cytoplasmic leakage, this binding can have a neutralizing effect on the biological properties of endotoxins25. The endotoxin of Gram-negative bacteria is the lipid A portion of LPS molecules, and colistin binds and neutralizes LPS. The significance of this mechanism for in vivo antimicrobial action, namely prevention of the endotoxin’s ability to induce shock through the release of cytokines, is not clear, because plasma endotoxin is immediately bound by LPS-binding protein, and the complex is quickly bound to cell-surface CD1442. More recently, the effects of slow-releasing colistin microspheres suggest that higher blood concentrations of colistin reduce the levels of endotoxin and cytokines in endotoxin-induced sepsis, and lead to decreased toxicity58. An alternative mechanism has recently been presented, demonstrating that polymyxins induce rapid cell death through hydroxyl radical production. Notably, it was demonstrated that this mechanism of killing occurs in multi-drug-resistant clinical isolates of A. baumannii and that this response is not induced in a polymyxin-resistant isolate59.

Mechanisms underlying colistin resistance

Unfortunately, colistin does not escape development of resistance. Although the incidence of colistin resistance is currently relatively rare worldwide, likely reflecting the uncommon use of colistin, colistin-resistant isolates have recently been identified in several GNB species such as A. baumannii, K. pneumoniae and P. aeruginosa60. It is important that the colistin-resistant clones do not spread to non-infected people since colistin is an important antibiotic for eradication of initial and intermittent P. aeruginosa colonization61.

Although the mechanism underlying resistance is unclear, resistance to the colistin as other polymyxins family has several molecular mechanisms that has been characterized in various bacterial species. It has been suggested resistance to this antibiotic is related to LPS modification via diverse routes. These include: (i) specific modification of outer membrane porins and reductions in the overall negative charge of the LPS, (ii) overexpression of efflux pump systems, (iii) overproduction of capsule polysaccharide62, and (iv) although enzymatic mechanisms of resistance haven’t been reported so far, strains of B. polymyxa are known to produce colistinase63.

Colistin resistance in Gram-negative bacteria is most commonly due to decreased binding to the bacterial outer membrane because of lipopolysaccharide remodeling that is caused by changes in PhoPQ and PmrAB, both two-component regulatory systems, which results in a less anionic lipid A, which stop or reduce this initial interaction with a different mechanism (Figure 4)64^,65. The pmr locus is an auto-regulated two-component signal transduction system, which in addition to a sensor kinase and response regulator, also includes an ethanolamine transferase. The ethanolamine transferase contributes to colistin resistance by adding ethanolamine moieties to the lipid A component of LPS, which reduces the negative charge of the bacterial membrane and thereby decreases binding of positively charged colistin66. The expression pattern of two-component regulatory systems in the presence of colistin was different from that in the absence of colistin: overall, pmrAB, pmrD, and phoPQ expressed the highest at the stationary phase in the absence of colistin, but the expression of pmrD and phoPQ decreased with the growth in the amount of colistin67. In many Gram-negative bacteria, acquired polymyxin resistance is most often mediated by replacement of lipid A by addition of 4-amino-4-deoxy-L-arabinose (L-Ara4N) and/or phosphoethanolamine (PEtn)55. This action requires the products of the ugd and pbg loci and ethanolamine, mediated by pmrC, therefore these modifications remove negative charges, lowering the affinity of LPS, thereby increasing resistance to polymyxins68. In addition, a non-specific mechanism for the tolerance of the metabolically active cells to colistin was shown to be up-regulation of the MexAB-OprM efflux pump69. According to the CLSI an MIC of ≤2 mg/L is susceptible and an MIC of ≥4 mg/L is resistant for Acinetobacter spp., which is different from P. aeruginosa (≥8 mg/L) and other non-Enterobacteriaceae70^,71.

Figure 4. PmrA/PmrB and PhoP/PhoQ two-component regulatory systems. (From Falagas et al.53).

Colistin resistance in A. baumannii

According to data colistin resistance is relatively rare in A. baumannii, and little is known about its mechanism. Moffatt et al. in their study reported spontaneously occurring, lipid A-deficient, Gram-negative bacterial mutants. Moreover, they provided evidence that loss of LPS can lead to polymyxin resistance, that loss of LPS production in A. baumannii can result from inactivation of a lipid A biosynthesis gene, lpxA, lpxC, or lpxD, and that colistin resistance via loss of lipid A can occur in clinical isolates55. These authors recently showed that insertion sequence ISAba11 movement can result in inactivation of the A. baumannii lipid A biosynthesis genes lpxA and lpxC, resulting in the complete loss of lipopolysaccharide production and high-level colistin resistance72. Loss or decreased expression of OmpW porin was also reported as a possible colistin resistance mechanism in the A. baumannii mutant73. Resistance can arise through mutations in the two component system pmrA and pmrB, in which the downstream target PmrC catalyzes the addition of PEtn to the lipid A component of LPS74. Park et al. also indicated that increased expression of the PmrAB system is essential for colistin resistance in A. baumannii but that amino acid alterations might be only partially responsible for resistance60. Although PmrAB and PhoPQ play an important role in colistin resistance in other bacteria such as S. enterica and P.aeruginosa, PhoPQ is absent in A. baumannii60^,75.

Colistin resistance in P. aeruginosa

Although a major route of adaptive polymyxin resistance involves the multi-component regulatory genes PmrAB, PhoPQ, ParRS and CprRS, which are known to be involved in polymyxin resistance in P. aeruginosa, the precise molecular details of these resistance mechanisms have also remained unclear, including a correlation of polymyxin susceptibility with amino acid alterations in these two-component systems76^,77. However, expression of the ParRS and CprRS genes was not consistent78. The PmrAB system responds to limiting Mg²⁺, and is affected by a phoQ, but not a phoP mutation. Inactivation of the pmrB sensor kinase and pmrA response regulator greatly decreased the expression of the operon encoding pmrA–pmrB while expression of the response regulator pmrA in trans resulted in increased activation suggesting that the pmrA–pmrB operon is autoregulated79. The PhoPQ system is a global regulatory system that autoregulates the oprH–phoP–phoQ operon under divalent cation-limiting conditions and in the presence of polyamines, which induce resistance to cationic peptides such as colistin80. Data from Lee et al. indicated that amino acid substitutions of PmrAB or PhoPQ do not have an immediate connection with decreased susceptibility of colistin in P. aeruginosa isolates, although activated expression of pmrAB and/or phoPQ resulting in overexpression of pmrH may be required for colistin resistance78. PmrAB directly controls the pmr HIJKLM operon, whose gene products are responsible for the synthesis of N4-aminoarabinose, which binds to lipid A moieties and neutralizes the negatively charged phospholipids, causing a change in the negatively charged cell membrane which leads to colistin resistance80.

Colistin resistance in Enterobacteriaceae

The expression of the L-Ara4N and PEtn transferases in E. coli and Salmonella enterica is also regulated by the two component regulatory system PmrA/PmrB, which sensing environmental properties such as pH, Fe³⁺ and Mg²⁺ level, as well as the presence of polymyxins, leading to altered expression of a set of genes involved in lipid A modification9^,55.

Salmonella typhimurium regulates mechanisms of resistance to cationic antimicrobial peptides through the two-component systems PhoP–PhoQ and PmrA–PmrB81. Polymyxin resistance is encoded by the PmrA–PmrB regulon, whose products modify the LPS core and lipid A regions with ethanolamine and add aminoarabinose to the 4 phosphate of lipid A81. Mutations in the pmrA locus of S. typhimurium confer an increase in resistance to polymixins. In addition, pmrA mutants survive better in human neutrophils, suggesting that this locus plays a role in virulence82. Polymyxin-resistant mutants of S. typhimurium and E. coli have a higher substitution of the ester-linked phosphate group in the lipid A portion of the LPS by 4-amino-4-deoxy-L-arabinose and show larger amounts of 2-aminoethanol esterifying phosphates in the core oligosaccharide83. The 4-aminoarabinose substitution is almost stoichiometric in strains of Proteus mirabilis, Chromobacterium violaceum, and Burkholderia cepacia that exhibit innate resistance to polymyxins83.

In KPC-producing K. pneumoniae (KPC-KP) it has been suggested that the mgrB alteration can be a common mechanism of colistin resistance in the clinical setting84. Comparative genomic analysis of a pair of sequential KPC-KP isolates from the same patient including a colistin-susceptible isolate (KKBO-1) and a colistin-resistant isolate (KKBO-4) selected after colistin exposure revealed that insertional inactivation of the mgrB gene, encoding a negative regulator of the PhoQ/PhoP signaling system, is a genetic mechanism for acquired colistin resistance85. A recent study confirmed the MgrB regulatory role in K. pneumoniae and was in agreement with the known association between upregulation of the PhoQ/PhoP system and activation of the pmr HFIJKLM operon, which eventually leads to resistance to polymyxins by modification of the lipopolysaccharide target.

Antibacterial combination therapy with colistin

The proportion of cells exhibiting heteroresistance from patients treated with colistin is significantly high86. Colistin heteroresistance, which has been identified in K. pneumoniae, A. baumannii, and P. aeruginosa, is thought to contribute to the rapid emergence of colistin resistance observed with monotherapy87. In clinical practice, it is frequently used as combination therapy in order to improve its antibacterial activity, despite the consequent increase in toxicity88. As simply increasing polymyxin dosage regimens is not an option for optimizing their PK/PD due to nephrotoxicity, combination therapy with other antibiotics has great potential to maximize the efficacy of polymyxins while minimizing emergence of resistance18. There have been contradictory reports about the effectiveness of combination therapy versus monotherapy and some studies have rejected it. However, the clinical cure rate and microbiological eradication rate were higher in the colistin combined therapy group than in the colistin monotherapy group (48.6% vs. 42.5% and 66.7% vs. 59.4%, respectively), although the differences were not statistically significant89.

Recently combination of colistin and a bacteriocin (nisin) showed a synergistic effect against Gram-negative bacteria and offers the possibility of eliminating the toxicity of this drug, as evidenced by the experiments carried out with mammalian cells90. In a study about in vitro activities of colistin combinations against P. aeruginosa isolated from the intensive care unit, synergistic activity of combinations of colistin-azithromycin, colistin-doxicycline and colistin-rifampicin was less than expected and a high percentage of indifferent results was observed91. Colistin resistance seems to promote the in vitro activity of unconventional colistin combination. Little impact was demonstrated on vancomycin, trimethoprim, or trimethoprim–sulfamethoxazole MIC values51.

In combination, synergy was observed with rifampicin, imipenem, arbekacin, aztreonam, piperacillin, and ciprofloxacin in strains, and according to data this combination is safe and effective. Maximum synergy was observed with colistin plus rifampicin (synergy, 80.0%; additive, 17.5%)94–94. For pneumonia especially, intranasal colistin with rifampicin may be beneficial not only for synergistic antibacterial activity, but also for blocking LPS92. This combination may have a role in the treatment of multi-drug-resistant K. pneumoniae and may possibly slow the selection of heteroresistant subpopulations during colistin therapy95. Other antimicrobial combinations with carbapenems, gentamicin, and tigecycline also showed variously synergistic results95. According to data in critically ill patients with carbapenem-resistant A. baumannii infections, clinical outcomes do not differ in patients treated with colistin plus vancomycin from those receiving colistin without vancomycin and this combination significantly increases the risk of renal failure96. The colistin/sulbactam combination therapy reported is promising in severe MDR A. baumannii VAP (ventilator-associated pneumonia). Although the difference was not statistically significant, clinical cure rates and bacteriological clearance rates were better in the combination group than colistin monotherapy97. Colistin-based combination therapy resulted in significantly higher microbiological eradication rates, relatively higher cure and survival rates, and lower in-hospital mortality compared to colistin monotherapy98.

Colistin combination therapy may be required to treat biofilm-associated infections. Colistin and doripenem in combination against biofilm-embedded and planktonic MDR P. aeruginosa, doripenem enhanced killing by colistin of biofilm-embedded cells in both carbapenem-susceptible and -resistant strains, and the combination minimized the emergence of colistin resistance99.

Adverse effects

It was reported in the old literature that the use of polymyxins was associated with a high incidence of toxicity, such as nephrotoxicity, neurotoxicity, and neuromuscular blockade, sometimes with fatal consequences. Therefore colistin use was limited in the late 1970s, except for treatment of patients with CF1^,100–102. However, recent studies showed that the incidence of nephrotoxicity is less common and severe compared to the old studies. In addition, neurotoxic effects of polymyxins are usually mild and resolve after prompt discontinuation of the antibiotics. Furthermore, cases of neuromuscular blockade and apnea have not been reported in the recent literature103.

Higher colistin doses, similar to those commonly used in the United States, led to a relatively high rate of nephrotoxicity104. Their nephrotoxicity may complicate the therapy or even require its discontinuation. It was reported that colistin monotherapy was not inferior to colistin combination therapy with respect to cure of the infection and nephrotoxicity105. It is notable that nephrotoxicity due to colistin treatment is more common in patients older than 60 years and is related to low initial glomerular filtration rate (GFR)106, so it is wise that patients receiving intravenous colistin should be monitored for nephrotoxicity.

Additionally, the use of polymyxins has been associated with the experience of several neurotoxic events, including dizziness, muscle weakness, facial and peripheral paresthesia, vertigo, confusion, ataxia and neuromuscular blockade, which can lead to respiratory failure or apnea, convulsions and coma107. Death was attributed to colistin therapy in less than 5% of cases58.

Prevalence of colistin resistance

As has been mentioned, nowadays the prevalence of colistin resistance is low, and it has been reported in a few studies. Increasing use of colistin for MDR-GNB infections has led to the emergence of colistin resistance in several countries worldwide. Prevalence may vary between regions and over time; however, some countries (e.g. Japan and South Africa) haven’t access to colistin and some areas of the world (e.g. Europe, Australia) have only the parenteral formulation of colistin (as CMS), whereas in other areas (e.g. USA, Brazil, Malaysia, and Singapore) clinicians can use either colistin or the polymyxin B parenteral formulation108. Despite the majority of the remaining reports presenting resistance rates under 10%, there are also some studies that report high resistance rates. It should be noted that these variations could exist due to differences in methodology. Some laboratories use disk diffusion, and polymyxins diffuse poorly due to high molecular weight. Hence, one could underestimate resistance when using disk diffusion53.

In a study by Gales et al. from the SENTRY antimicrobial surveillance program, in worldwide collection of Gram-negative pathogens reported resistance to the polymyxins remained stable among organisms tested except for Klebsiella spp. isolates collected from the Asia-Pacific and Latin American regions. Indeed, only 0.4% of P. aeruginosa, 0.9% of Acinetobacter spp. and 1.5% of Klebsiella spp. were resistant to colistin43. Another study in the SENTRY antimicrobial surveillance program in 2008 examined the susceptibility to colistin of multi-drug-resistant A. baumannii from the Western Pacific region. It revealed that all the isolates were highly multi-resistant, and colistin MICs were 0.5–2 mg/L except one isolate which had an MIC of 128 mg/L109.

In addition to colistin resistance, heteroresistance has been identified, which has been observed in several Gram-negative pathogens4^,86. Heteroresistance is broadly defined as the presence of an antibiotic-resistant subset of microbes within a larger population that is susceptible to the antibiotic. Heteroresistance can complicate assessment of the MIC to a specific antibiotic and may promote resistance to antibiotics in vivo, thereby affecting diagnostic tests and patient treatment110. However, because of a lack of uniform standards to determine heteroresistance, the rates from different regions varied greatly.

American countries

Most reports from the USA represent resistance rates for both P. aeruginosa and A. baumannii as less than 5.5%111^,112. However, studies of resistance in K. pneumoniae report higher rates53. Development of resistance to colistin in multi-drug-resistant P aeruginosa has not been observed in Denmark despite the use of nebulized colistimethate sodium in combination with oral ciprofloxacin19. The in vitro activity of colistin evaluated against Gram-negative bacilli from patients in Canadian hospitals during 2007–2008 showed colistin was active in vitro (MIC₉₀, ≤2 mg/L) against a variety of clinically important Gram-negative bacilli, including E. coli, Klebsiella spp., Enterobacter spp., A. baumannii, and P. aeruginosa; also all MDR P. aeruginosa clinical isolates evaluated remained susceptible to colistin113. To identify colistin-resistant A. baumannii isolates in an intensive care unit from Argentina, Colistin heteroresistance was observed in 46 (4%) of these isolates and the majority belonged to clones previously identified as I and III114. Other countries in South America, including Brazil and Chile, report resistance rates of up to 9% for P. aeruginosa and A. baumannii53.

African countries

African reports are scarce. Studies from Nigeria and South Africa reported that resistance rates were less than 10% and fully susceptible to colistin, respectively115^,116. Mezghani Maalej et al. from Tunisia in a retrospective study isolated 121 strains of colistin resistant Enterobacteriaceae from 93 patients. The study reported the rate of resistance to colistin ranged from 0.09% for E. coli to 1.2% for K. pneumoniae, and 1.5% for E. cloacae117. However, from Zimbabwe cumulative data for P. aeruginosa from two different hospitals reported resistance rates of 53%118.

European countries

There are different reports about the resistance rate in Europe. In the UK a survey of CF patients reported that colistin was clearly the most active antimicrobial tested, with 97% of isolates susceptible to <4 mg/L with susceptibility testing by E-test, and 3.1% of P. aeruginosa isolates were resistant to colistin119.

Arroya et al. in 2005 compared the E-test to the broth microdilution method for testing the colistin susceptibility of 115 clinical isolates of A. baumannii. In their study from Spain twenty-two (19.1%) strains were resistant to colistin120. From Romania the resistance rate of E. coli and Klebsiella spp. were reported to be 11% and 17%, respectively121.

Capone et al. during 2010–2011 performed a prospective study of patients with carbapenem-resistant K. pneumoniae isolation, hospitalized in nine hospitals in Rome, Italy. They observed a high rate of resistance to colistin that is independently associated with worse outcome; 36.1% of strains were also resistant to colistin122. They suggested that the increased use of this drug during recent years, especially as monotherapy, could be the cause of this. Mammina et al. also described polyclonal spread of colistin-resistant K. pneumoniae in an acute general hospital in Italy. In their study fifty-two isolates had MICs for colistin of 6–128 mg/L, carried bla_KPC3 and were attributed to sequence type ST258123.

In 2003, at least nine isolates were reported non-susceptible to colistin in Greece124. A retrospective cohort study of colonization and infection by colistin-resistant Gram-negative bacteria in critically ill patients from Athens (Greece) during 2003–2006 revealed that 52% of patients were colonized by colistin-resistant Gram-negative bacteria, of whom 20% were colonized by K. pneumoniae isolates and 34% were colonized by intrinsically resistant to colistin (CIR) enterobacteriaceae125. Another study from Greece during 2005–2008 reported that 10.5% of Klebsiella pneumonia isolates were non-susceptible to colistin and among the imipenem-resistant isolates 20% were also resistant to colistin although no resistance was observed during 1996–1998126. It was reported that colistin resistance had risen at an alarming rate, from 1% in 2005 to 19% in 2008. In Greece’s study published in 2010, all isolates were resistant to carbapenems, β-lactams, ciprofloxacin, aminoglycosides and colistin, but intermediately susceptible to tigecycline and susceptible to gentamicin127.

Resistance to this drug in K. pneumoniae has been extensively described in Hungary. During 2008–2009, in the north-eastern region of Hungary, nine K. pneumoniae isolates showed non-susceptibility to carbapenems, and of these, eight isolates were highly resistant to colistin, which isolates carried bla_KPC-2, bla_SHV-12, bla_TEM-1 and bla_SHV-11128. Nowadays epidemiological studies carried out in European countries also observe high colistin use in animal health129.

Asian countries

Studies from the Asia-Pacific region show that resistance to colistin is common in Enterobacter spp. and has also been detected in Klebsiella spp., and frequently manifests a heteroresistance pattern. In a study resistance to colistin was detected in 0.3% Klebsiella spp. and 21% Enterobacter spp.; however, resistance was seen in all countries except Singapore, ranging from 13.8% (India) to 50% (Philippines)130. In a study from Singapore, 30% of P. aeruginosa isolates were resistant to colistin. In the same study all Acinetobacter spp. and Escherichia coli were susceptible to colistin, and resistance was detected predominantly in S. maltophilia and P. aeruginosa, but was also present in Enterobacter spp. and Klebsiella spp.40. A surveillance study of carbapenem-resistant Enterobacteriaceae isolates from Shanghai (China)131 noticed independent emergence of colistin resistance in KPC-producing carbapenem-resistant Enterobacteriaceae (CRE) isolates without clinical treatment with colistin. In a recent study from China, susceptibility rates for colistin were 92.7%. Although colistin was found to be most active against CRE isolates, four isolates showed high resistance to colistin, with MICs of >64 mg/L for three isolates and 4 mg/L for one isolate131. Another study from China reported that of 112 non-repetitive clinical isolates of A. baumannii–A.calcoaceticus complex, 80% were resistant to a variety of structurally unrelated antimicrobials and resistance to carbapenems occurred in 8% of the isolates, although all isolates were susceptible to polymyxin132.

Iranian data revealed that the rate of colistin resistant A. baumannii with the E-test method was 11.6%. So, as the frequency of resistance to colistin is low, it can be used as an easily available drug for treatment of MDR A. baumannii strains, which are susceptible to colistin133. But in another study on 91 A. baumannii isolates from patients in tertiary intensive care units of three university hospitals in north, central, and south Iran, the drug resistance pattern showed that 14.2%, 20%, and 77% of the A. baumannii isolates were resistant to colistin, tigecycline, and rifampicin, respectively134.

According to recent data, high colistin resistance rates in Acinetobacter spp. were reported in Korean hospitals. Ko et al. identified that 18.1% and 27.9% of A. baumannii strains were resistant to polymyxin B and colistin, respectively, belonging to subgroups II and III in Korea, on the basis of rpoB gene analysis135. Multilocus sequence typing (MLST) analysis indicated that most colistin-resistant Acinetobacter spp. isolates from Korean hospitals arose independently136. In another study conducted in South Korean hospitals between 2006 and 2007 on patients with bacteremia, 6.8% of K. pneumoniae isolates were resistant to colistin137. In a recent study the clonality of colistin-resistant K. pneumoniae (CRKP) isolates was assessed by MLST. MLST also showed that CRKP isolates were nonclonal, with colistin resistance in K. pneumoniae occurring independently and not by clonal spreading. Studies conducted in Turkey showed all A. baumannii strains were 100% susceptible to colistin102^,138–140. However, a study on A. baumannii from Turkey reported resistance rates of 27.5%141. Although colistin appears to be a good choice, adverse reactions and unavailability of colistin limit its wide usage in Turkey140.

Studies from Thailand reported that colistin is a good option for treatment of MDR P. aeruginosa142^,143. In one study colistin had good efficacy against MDR P. aeruginosa with a 98% susceptibility rate and MIC₅₀ and MIC₉₀ for colistin reported to be 1.0 and 1.5 mg/L, respectively, except that two MDR P. aeruginosa isolates were resistant to colistin with MICs of 3 and 12 mg/L, respectively143. Moreover, 7.3% colistin resistance for K. pneumonia was reported from Thailand144. In India a high prevalence of resistance was observed in both P. aeruginosa and Acinetobacter isolates to most antibiotics except to colistin, including carbapenems, with rates ranging from 22.1–57.9% and 42–94.0%, respectively145. However, 38.3% of P. aeruginosa in 2008 has shown resistant to colistin146. In this country it was also reported that all Metallo-beta-lactamase (MBL)-producing P. aeruginosa producers were susceptible to colistin with MIC ranging from 0.5–0.032 mg/L147. In another study in the case of Acinetobacter spp. which also conducted in North India, 3.5% (from 224 isolates) were resistant to both tigecycline and colistin148. In Israel in a population of injured military personnel returning from Iraq and Afghanistan, the most active agents (>95% of isolates susceptible) were colistin, polymyxin B, and minocycline149^,150.

Conclusion

Colistin is a cationic antimicrobial peptide, with a narrow spectrum of activity mainly against Gram-negative bacteria, that returned to clinical practice due to the lack of options for the treatment of MDR-GNB infections. However, its adverse effects may complicate therapy or even require its discontinuation. It acts by disrupting the bacterial membranes through anionic displacement of stabilizing magnesium and calcium, resulting in leakage of cell contents and eventually in cellular death. Increasing use of colistin for treatment of infections caused by these bacteria has led to the emergence of colistin resistance in several countries worldwide. The highest resistance rate was reported in Asia (especially Korea and Singapore), followed by Europe (especially Greece) and America, where colistin resistance rates are continually increasing. In summary, our review demonstrated that colistin is highly useful as a preferred alternative agent and it was at least as effective as or even more effective than β-lactams, quinolones, and aminoglycosides in the treatment of MDR-GNB patients.

Transparency

Declaration of funding

This study was supported by a grant from the Drug Applied Research Center (Tabriz University of Medical Sciences) with Grant No. 93007782.

Declaration of financial/other relationships

A.Z.B. and H.S.K. have disclosed that they were recipients of the research grant named above.

CMRO peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

The authors thank all staff of educational hospitals for cooperation in antibiotic stewardships of Tabriz University of Medical Science. They also thank Dr. Hossein Navidinia for his helpful comments on manuscript.