Colistin is effective against

Increasing incidence of resistance of gram-negative bacteria against even newer antibiotic including carbapenem has generated interest in the old antibiotic colistin, which are being used as salvage therapy in the treatment of multidrug resistant infection. Colistin has excellent bactericidal activity against most gram-negative bacilli. It has shown persist level in the liver, kidney, heart, and muscle; while it is poorly distributed to the bones, cerebrospinal fluid, lung parenchyma, and pleural cavity. Being an old drug, colistin was never gone through the drug development process needed for compliance with competent regulatory authorities that resulted in very much limited understanding of pharmacokinetic [PK] and pharmacodynamic [PD] parameters, such as C max/MIC ratio, AUC/MIC and T > MIC that could predict the efficacy of colistin. In available PK/PD studies of colistin, mean maximum serum concentration [C max] of colistin were found just above the MIC breakpoint at steady states that would most probably lead to suboptimal for killing the bacteria, even at dosages of 3.0 million international units [MIU] i.e., 240mg of colistimethate sodium [CMS] intravenously every 8h. These finding stresses to use high loading as well as high maintenance dose of intravenous colistin. It is not only suboptimal plasma concentration of colistin but also poor lung tissue concentration, which has been demonstrated in recent studies, poses major concern in using intravenous colistin. Combination therapy mainly with carbapenems shows synergistic effect. In recent studies, inhaled colistin has been found promising in treatment of lung infection due to MDR gram-negative bacteria. New evidence shows less toxicity than previously reported.

Worldwide, there are growing threats to modern medicine from the emergence of multidrug resistant [MDR] bacteria causing nosocomial infection. This coupled with marked decline in the discovery and development of novel antibiotics especially against gram-negative bacteria in the last two decades leads to critical challenge to clinicians. MDR bacteria are usually defined as when it is resistant to three or more group of antibiotics. In clinical practice, antibiotics commonly used in treating gram-negative infection are penicillin, cephalosporin, carbapenem, monobactem, quinolone, and aminoglycoside. Increasing incidence of resistance of gram-negative bacteria against even newer antibiotic including carbapenem has been reported in many countries. In the last few years, with a paucity of available antibiotic option, interest has been renewed in the old antibiotic, polymyxins, as salvage therapy in the management of infections caused by MDR gram-negative pathogens including Pseudomonas, Acinetobacter, Klebsiella, and Enterobacter species [1-3].

Polymyxins are group of cationic polypeptide antibiotics, having five different compounds [polymyxin A-E] [4]. In 1949, colistin [Polymyxin E] was first time isolated from Bacillus polymyxa var. colistinus by Koyama Y. and colleagues in Japan [Table1] [5]. Only polymyxin E [Colistin] and polymyxin B have been used in clinical practice since their discovery [1,2]. Both, colistin and polymyxib B, are produced by the soil bacterium Bacillus spp., and they differ in structure only by one amino acid at position 6 [-Leu in colistin, while -Phe in polymyxin B] [4].

Clinical experience is more with colistin in comparison to polymyxin B due to its wider use. After introduction in clinical practice, uses of parenteral colistin were gradually waned worldwide within two decades due to reported severe toxicities, while its topical uses continued [2-4]. So, era of colistin as intravenous use could be divided in three phases: 19501970: against gram-negative infection; 19902000: for MDR gram-negative pathogens in cystic fibrosis; and 2000 onwards: nosocomial infection due to MDR gram-negative pathogens. In this presented descriptive review, we looked in published literature for the use of colistin in lung infection in critically ill patients without cystic fibrosis.

Current issues in colistin therapy

Colistin, itself is composed of mixture of closely related components, mainly colistin A [polymyxin E1] and colistin B [polymyxin E2], which are acylated by [S]-6-methyloctanoic acid and [S]-6-methylheptanoic acid, respectively [Figure1] [4]. Each molecule has a cationic polypeptide ring with a lipophilic fatty acid chain.

Spectrum of activity

Colistin has excellent bactericidal activity against most gram-negative bacilli, while no activity against all gram-positive and most anaerobes [except few like Prevotella spp., Fusobacterium spp.] organisms [Table2] [6]. Colistin has also been reported to be potentially active against several Mycobacterium spp. [7].

Mechanism of action and resistance

The targets site for colistin is lipopolysaccharide [LPS] component of the outer membrane of gram-negative bacteria. In there, colistin interacts with the lipid A component of the LPS, displacing the calcium and magnesium bridges that stabilize the LPS, leading to permeabilizing the bacterial outer membrane that can be described as detergent-like mechanism of action [4]. Subsequently, colistin inserted through these cracks in the outer membrane of bacteria causes self-promoted uptake as well as disruption of the integrity of the inner membrane also that leads to bacterial killing [1,4].

By another postulated mechanism, polycationic polymyxin binds to anionic phospholipid; this contact promotes lipid exchange between inner and outer membrane results in the loss of phospholipid composition leading to an osmotic imbalance resulting in lytic cell death [4].

The majority of mechanisms of resistance to colistin are based on modifications to the lipid A portion of LPS of gram-negative bacteria, like Pseudomonas aeruginosa and Escherichia coli, that reduces its net negative charge resulting in less electrostatic interaction with positively charged colistin molecule [1,4]. While in Klebsiella pneumoniae, the mechanism of resistance to colistin is because of increased production of capsule polysaccharide [8].

Susceptibility test and heteroresistance

Antimicrobial susceptibility testing for colistin can be performed using either diffusion [disc diffusion, prediffusion test, or epsilometer test [E-test] or dilution techniques [broth microdilution or agar dilution] [6]. Colistin sulfate is recommended in susceptibility testing instead of colistimethate sodium [CMS] [2,6]. The reason for not using CMS in susceptibility testing, which is used clinically [see below], is an inactive pro-drug and also undergo variable hydrolysis resulting in differential killing characteristics and varying susceptibility report. There is another issue that colistin has, that because of its large molecule, it diffuses inadequately into the medium which makes disc diffusion test as a poor performer in susceptibility test. Among all these methods, broth microdilution and agar dilution methods are considered as reference methods [6,9]. Alternatively, E-test method is a reliable, easy to perform, and less time-consuming [6].

The breakpoints for colistin susceptibility are defined differently by two main societies: as per US Clinical and Laboratory Standards Institution [CLSI], 2mg/L as the susceptibility breakpoint and >2mg/L as the resistance breakpoint, while as per British Society for Antimicrobial Chemotherapy [BSAC], 4mg/L as susceptible and 8mg/L as resistant [2].

In a population analysis, profiles of colistin-susceptible clinical isolates Acinetobacter baumannii showed that subpopulation  MIC that could predict the efficacy of colistin [2,11]. Recent studies started to reveal about PK/PD of colistin and also demonstrate that it has poor postantibiotic effect [PAE] [11].

In PK/PD studies of colistin, mean maximum serum concentration [C max] of colistin were found just above the MIC breakpoint of 2mg/L [Clinical and Laboratory Standards Institute] at steady states that would most probably lead to suboptimal for killing the bacteria, even at dosages of 3.0 MIU i.e., 240mg of CMS intravenously every 8h [12-14]. Steady state concentration was achieved at least after 48h of starting the intravenous colistin [12]. In a population of critically ill patients where pharmacokinetic analysis of colistin was done after intravenous administration, Plachouras et al. predicted that plasma colistin concentration on the basis of the model developed for different dosage regimen revealed that loading dose of 9 or 12 MIU of CMS as infusion of 15min or 2h with maintenance dose of 4.5 MIU CMS every 12h achieves rapid [4μg/ml killed P. aeruginosa, whereas only 40% of samples containing colistin 8 was achieved against Acinetobacter, not for Pseudomonas. Bergen et al. examined the PK/PD relationship of colistin against P. aeruginosa and found that colistin efficacy against P. aeruginosa was best correlated with the AUC/MIC ratio of total and unbound colistin rather than the C max/MIC ratio [16].

It is not only the optimal plasma concentration but the site of infection also defines the optimal antimicrobial therapy. Interestingly, in a study, Imberti et al. found undetectable level of colistin in broncho-alveolar lavage fluid of critically ill patients at steady state mean plasma colistin maximum concentration of 2.21μg/ml after the intravenous administration of CMS 2 MIU every 8h for at least 2days [17].

Use in patients with renal dysfunction

Despite colistin is eliminated by non-renal route, dose adjustment for renal dysfunction is necessary, as there is reduced renal dependent elimination of its pro-drug CMS which ultimately lead to increased colistin level [2,11]. Despite scarcity of studies of PK/PD of colistin in patients with renal failure, recent recommended doses are [2]:

• Serum creatinine level 1.31.5mg/dl: 2 MIU [160mg] of CMS every 8h

• Serum creatinine level 1.62.5mg/dl: 2 MIU [160mg] of CMS every 12h

• Serum creatinine level 2.6mg/dl: 2 MIU [160mg] of CMS every 24h

Patient on renal replacement therapy:

• 2 MIU [160mg] of CMS after each hemodialysis

• 2 MIU [160mg] of CMS daily during peritoneal dialysis

In a study about pharmacokinetics of colistin in critically ill patients receiving continuous venovenous hemodiafiltration [CRRT], Karvanen M et al. suggested that dose regimen of 2 MIU CMS every 8h is inadequate [18]. In another recently published study, Honore et al. postulate from their experience that patients undergoing CRRT may receive substantially higher doses of colistin, i.e., high loading dose followed by a maintenance dose of up to 4.5 MIU every 8h because the drug is continuously filtered and also significantly adsorbed in the bulk of the dialysis membrane [19].

Inhaled colistin

For aerosol inhalation purpose, both forms of colistin [colistin sulfate and CMS] could be used [2,3]. Colistin, in the same intravenous formulation, is dissolved in 46ml of normal saline or sterile water to deliver through nebulizer, though this is not approved by the Food and Drug Administration [FDA] [20,21]. There are few studies on the PK of colistin after inhalation. Ratjen et al. evaluated the colistin pharmacokinetics postinhalation in patients with cystic fibrosis [22]. This study finds that a single dose of CMS [2 MIU] achieve significant higher drug concentration in the sputum even after 12h with low level in serum and urine. In another experimental study done by Lu et al. where pneumonia caused by P. aeruginosa in piglets and CMS was administered either by nebulization every 12h or intravenous every 8h, lung tissue concentration of colistin was measured [23]. Colistin was found undetected in the lung tissue after intravenous infusion, while after nebulization, peak lung tissue concentrations were significantly higher in the lung segments [more in mild pneumonia segments and less in severe pneumonia area, median 10.0 versus 1.2μg/g].

As per drug package insert information, the recommended doses of colistin when given by inhalation are as below [24]:

• Body weight 40kg: 1.0 MIU [80mg] of CMS every 12h

• For recurrent or severe pulmonary infection: 2.0 MIU [160mg] of CMS every 8h

Optimal inhalation therapy also requires consideration of several factors like the type of nebulizer, aerosol particle size, position of patient, severity of airway obstruction, etc. In a mechanically ventilated patient, there are factors more than this, i.e., artificial airway size, humidity, gas density, tidal volume, nebulization cycling during inspiration versus continuous, etc., which may affect drug delivery at the target site [25].

Combined therapy with other antibiotics and synergism

In clinical practice, colistin is frequently used as combination therapy, though there is scarcity of data that whether combination therapy is superior to monotherapy. One of the reasons for limited studies may be using this drug in treating MDR pathogens that deterred the clinician from providing colistin as a monotherapy. In a recent review by Petrosillo et al. on colistin monotherapy versus combination therapy in animal and clinical studies, they revealed that synergistic effect was detected in all the nine studies examining the combination of colistin and rifampicin, whereas carbapenems exhibited a synergistic effect in two out of three studies [26]. Colistin combined with tigecycline did not show good synergistic action [27]. Considerable synergy was found between levofloxacin and colistin in a recently published study [28]. Pankuch et al. studied time killing synergy with subinhibitory concentration of meropenem and colistin and found significant synergy against A. baumannii at 6, 12, and 24h, while there was no significant synergy against P. aeruginosa at any point of time [29]. In another in vitro study by Souli et al., the combination of imipenem [irrespective of MIC] and colistin against K. pneumoniae showed synergistic activity against isolates susceptible either to both agents or to colistin, while rarely synergistic against non-colistin-susceptible strains [30]. A recently published study by Leu et al. showed that the synergy of imipenem and colistin against imipenem-non-susceptible multi-drug resistant A. baumannii was significantly better for the colistin concentration at 1mg/L than that at 0.5mg/L [31].

Toxicity

Two types of toxicity, namely, nephrotoxicity and neurotoxicity has been reported with the uses of colistin. A systematic review of the toxicity of polymyxin revealed that in the old literature, incidences of both toxicities were reported to be considerably high, while new evidence shows less toxicity than previously reported [32]. The observed nephrotoxicity was as high as 50% in old studies versus 15%25% in recent studies, although the definition of nephrotoxicity was not standardized between the studies [32]. Renal toxicity of colistin has been described as dose dependent and may be partly due to their D-amino acid content and fatty acid component that leads to acute tubular necrosis. Renal toxicity from colistin has been found apparently less than the use of polymyxin B [32]. Age [more than 60years] has also been found as a significant risk for colistin-induced nephrotoxicity [33]. The concurrent administration of nephrotoxic drugs, hypovolemia or shock, and severity of illness may increase the likelihood of development of acute kidney injury [AKI].

In a recent preliminary study on high-dose colistin [9 MIU twice-daily], administration in critically ill patients found that AKI developed during 18% of treatment courses did not require renal replacement therapy [RRT] and subsided within 10days from colistin discontinuation [34]. Also, there was no correlation between variations in serum creatinine level and daily and cumulative doses of colistin or duration of colistin treatment. In another study, Dewan et al. also reported AKI in 16% of patients in whom high dose, extended interval colistin [9 MIU stat followed by 4.5 MIU 12 hourly] was used, while no patient required RRT [35].

Overall incidence of neurotoxicity related to colistin use is less than the nephrotoxicity. Earlier studies reported paresthesias in about one forth of patients receiving colistin, with few case reports of neuromuscular blockade or apnea, while recent studies did not reported any significant neurotoxicity [32,36]. Neurotoxicity is also dose dependent and may be triggered by the presence of risk factors like the presence of hypoxia, co-administration of muscle relaxant, narcotics, sedatives, or steroids [32,37,38].

Colistin aerosol inhalation therapy is generally well tolerated with few reported side effects like throat irritation, cough, and bronchospasm, which may be because of the presence of excipients/preservatives [20].

Colistin for lung infection and clinical outcome

In the last decade, many studies published the use of colistin in the treatment of lung infection in patients without cystic fibrosis, with conflicting results. Most of the studies have population of adult ICU patients with MDR VAP, and types of study were single-arm [Table4] [11,39-75]. In a systematic review and meta-regression to know the efficacy of colistin for the treatment of VAP, Florescu et al. analyzed 6 controlled studies [359 patients] and single-arm analysis from 14 studies [437 patients] [76]. Among the six two-arm studies [three prospective including one RCT and three retrospective studies], which were unblinded, the mean duration of intravenous colistin was 11.4days at a mean dose of 252.5mg/70kg/day and aerosolized colistin was administered for a mean 9.25days at a mean dose of 355mg/70kg/day. While, in single-arm studies, intravenous colistin was used for a mean of 15days at a mean dose of 209.5mg/70kg/day and aerosolized form was given for a mean of 14days at a mean dose of 80mg/70kg/day. In the meta-analysis of six controlled studies, there was no significant difference for overall clinical response between colistin and control groups [OR 1.14 [95% CI 0.741.77; p = 0.56]; I2 = 0%; Q = 4.98 [p = 0.42]], even after controlling for concomitant antibiotic treatment, or for the dose of intravenous colistin. Though, there was a trend for better microbiological response. Also, there was no significant difference between the colistin and control groups with respect to ICU, hospital, or 28-day mortality. Neither, incidence of nephrotoxicity was different in both groups; but one study reported higher incidence for respiratory toxicity [55]. In the meta-analysis of single-arm studies, there was favorable clinical response with colistin found [95% CI 0.640.80; Q = 55.3; p