Efficacy of High-dose Nebulized Colistin in Ventilator-associated Pneumonia Caused by Multidrug-resistant Pseudomonas aeruginosa  and Acinetobacter baumannii 

• Colistin is often the only active antibiotic for multidrug-resistant (MDR) gram negative pathogens

• Using 5 million IU every 8 h of nebulized colistin in patients with ventilator-associated pneumonia caused by MDR organisms or intravenous β-lactams to patients with sensitive organisms, the clinical cure rate was similar in patients with sensitive (66%) or MDR (67%) P. aeruginosa and A. baumannii 

Ventilator-associated pneumonia (VAP) caused by Pseudomonas aeruginosa  and Acinetobacter baumannii  is characterized by high rate of recurrence and frequent selection of new resistance to antibiotics despite adequate initial antimicrobial therapy. The decreased susceptibility of these strains to antibiotics has become a major health problem worldwide. Nowadays, few antimicrobials are available to treat Gram-negative multidrug-resistant VAP, and often, colistin remains the only active agent.1 

Although synergistic antibiotic activity with colistin combined to rifampicin or carbapenems has been shown in vitro ,2,3the superiority of such combination therapy has never been demonstrated in patients.4–6Intravenous colistin monotherapy is therefore often used as salvage therapy in the treatment of patients with VAP caused by multidrug-resistant Gram-negative pathogens susceptible only to colistin.7–9Because of its poor lung tissue penetration,10,11the effectiveness of intravenous administration is highly questionable. Both success and failure of the treatment have been reported in a few observational, retrospective, or uncontrolled series. Another major concern regarding intravenous colistin monotherapy is the emergency of resistance.12,13The risk of acquisition of colistin resistance could be even higher if its lung tissue concentration is insufficient compared with minimal inhibitory concentration (MIC).14 

Nebulization of antibiotics offers the possibility of generating high drug concentrations at the site of infection.15–17In a previous experimental study, nebulized colistin monotherapy resulted in high lung deposition and bactericidal effects in ventilated piglets with inoculation pneumonia caused by P. aeruginosa  intermediate to ceftazidime.10A recent clinical study performed in patients with VAP reported that combined nebulization of ceftazidime and amikacin was effective against P. aeruginosa  intermediate to ceftazidime and/or amikacin and may prevent per-treatment acquisition of antibiotic resistance.18 

The objective of the study was to evaluate the efficacy of high-dose nebulized colistin for treating VAP caused by multidrug-resistant P. aeruginosa  and Acinetobacter baumannii . A group of patients with VAP caused by susceptible P. aeruginosa  or A. baumannii  treated with conventional intravenous antibiotics served as controls. The second objective was to assess the risk of developing colistin resistance after administration of nebulized colistin in patients with recurrent VAP.

This prospective, observational study was conducted from January 1, 2006, to December 31, 2010, in the 26-bed multidisciplinary intensive care unit (ICU) of La Pitié-Salpêtrière Hospital. The institutional review board of La Pitié-Salpêtrière Hospital, Paris, France, approved the study protocol. Because nothing more than routine diagnostic tests, monitoring, and treatment was performed during the study, informed consent from patients or their relatives was waived.

The eligible criteria for the study were age older than 18 yr, mechanical ventilation required for more than 48 h, and VAP caused by P. aeruginosa  or A. baumannii.  VAP was defined as the presence of new and persistent infiltrates on chest radiography and bedside lung ultrasound highly evocative of lung infection19,20associated with one of the following clinical features: (1) temperature ≥ 38.4°C or < 36.5°C; (2) leukocyte count > 11.10³/ml; and (3) purulent bronchial secretions. P. aeruginosa or A. baumannii  was confirmed in lower respiratory tract specimens sampled either by fiberoptic bronchoscopy with nonprotected bronchoalveolar lavage or protected mini-bronchoalveolar lavage.21A positive sample was defined as more than or equal to 10⁴colony-forming unit (CFU)/ml for nonprotected bronchoalveolar lavage and more than or equal to 10³CFU/ml for protected mini-bronchoalveolar lavage. Exclusion criteria were (1) severe immunosuppression, defined as leukocyte counts less than 1000 cells/ml or neutrophils less than 500 cells/ml; (2) extrapulmonary infection such as bacteremia, urinary infection, peritonitis, catheter-related infection, mediastinitis, meningitis, endocarditis, or skin infections; (3) patients treated with intravenous colistin; (4) patients treated with a combination of nebulized colistin and intravenous antibiotics other than aminoglycosides; and (5) allergy to colistin. Before obtaining the result of the bronchoalveolar lavage culture, each patient included in the study received either an empirical antimicrobial therapy based on the bacterial ecology of the ICU during a mean time of 1.6 ± 0.5 days or an antimicrobial therapy for treating the previous episode of VAP.

Two groups of patients were distinguished from the cohort patients: the sensitive strain group including patients with VAP caused by A. baumannii  or P. aeruginosa  susceptible to β-lactams and the multidrug-resistant strain group including patients with VAP caused by P. aeruginosa  or A. baumannii  resistant to all β-lactams and susceptible to colistin. When the first episode of VAP was caused by susceptible P. aeruginosa or A. baumannii  and followed by a second episode caused by multidrug-resistant P. aeruginosa or A. baumannii,  patients were included in the multidrug-resistant strain group rather than in the sensitive strain group. Patients’ flowchart is shown in figure 1.

Patients in the sensitive strain group received for 14 days intravenous β-lactam (ticarcillin/piperacillin, ceftazidime, or imipenem) combined either with aminoglycoside (78% of patients) or quinolone (22% of patients) for 3 days. This 14-day treatment regimen was selected to avoid a high pulmonary infection recurrence rate.22 

Patients in the multidrug-resistant strain group received an aerosol of 5 million international units (IU) (400 mg) of colistimethate every 8 h for 7–19 days. The duration of aerosol was maintained for 14 days or until successful weaning from mechanical ventilation. After extubation, aerosol of antibiotics cannot be properly delivered through the natural airways during spontaneous breathing because of inspiratory flow turbulence precluding reaching the deep lung. Aerosolized colistimethate dose was calculated according to a 40% extrapulmonary deposition as previously described.10Therefore, the resulting fraction of colistimethate reaching the respiratory tract was 60% of the initial dose placed in the nebulizer chamber, representing a daily dose equivalent to 3 million IU delivered to the respiratory tract every 8 h.

Colistin plasma concentrations were measured at day 2 and day 3 using high-performance liquid chromatography method.23,24Peak and trough serum concentrations were measured 30 min after nebulization and immediately before the next nebulization, respectively. All blood samples were immediately centrifuged for 15 min (4,000g ) at 4°C, and plasma samples were stored at −40°C for later analysis.

Serum creatinine was recorded on days 0, 3, 7, and 14 in both groups. Colistin-induced renal function impairment was defined as an increase in serum creatinine level more than or equal to 1.5 times the pretreatment value.25,26 

In the multidrug-resistant strain group, nebulization was performed with a vibrating plate nebulizer (Aeroneb Pro^®, Aerogen Nektar Corporation, Galway, Ireland) positioned on the inspiratory limb 10 cm proximal to the Y-piece.27After inserting 5 million IU of colistimethate powder diluted in 10 ml of sterile water into the nebulizer chamber, each nebulization was delivered over 60 min. Specific ventilator settings were used during the nebulization period to reduce flow turbulences and thereby extrapulmonary deposition.17They included removal of heat and moisture exchanger or conventional humidifier, volume controlled mode, administration of constant inspiratory flow, respiratory rate of 12 breaths/min, inspiratory–expiratory ratio of 50%, tidal volume of 8 ml/kg, and an end-inspiratory pause representing 20% of the duty cycle. During the nebulization period, expired aerosolized particles were collected in a filter with pore size equal to 0.2 µm positioned on the distal part of the expiratory limb (Hygrobac; Mallinckrodt Medical, Mirandola, Italy). Strict coordination between the patient and the ventilator was demanded to avoid inspiratory turbulences and optimize distal lung deposition of aerosolized particles. In case of discoordination, 2 mg/kg of propofol was infused. After each aerosol, the filter was removed and heat and moisture exchanger or conventional humidifier repositioned. To standardize the procedure of aerosol administration, a checklist form was completed by the nurse in charge of the patient as previously described.18 

Clinical responses were classified at the end of treatment at day 14 by independent physicians as cure, persisting VAP, recurrence, and superinfection. Cure of VAP was defined as resolution of clinical and biological signs of infection, clinical pulmonary infection score (CPIS) less than 6 and negative culture of lower respiratory tract specimens if available. Persisting VAP was defined as lack of improvement of clinical and biological signs of infection, CPIS greater than 6, and significant concentrations of P. aeruginosa  and A. baumannii  persisting in the lower respiratory tract. Recurrence was defined as initial cure of VAP with antimicrobial treatment at day 14 followed by the reappearance of clinical and biological signs of infection, CPIS greater than 6, and significant concentrations of P. aeruginosa  and A. baumannii  in lower respiratory tract specimens. Superinfection was defined as reappearance of VAP caused by pathogens other than P. aeruginosa  or A. baumannii  isolated from lower respiratory tract specimens.18 

All causative microorganisms were identified using routine microbiological methods. The disk diffusion method was used for antibiotic susceptibility testing, except for colistin. Mueller–Hinton agar and disks of antibiotics were purchased from Sanofi Diagnostics Pasteur (Marne la Coquette, France) and were used according to the guidelines of the Antibiogram Committee of the French Society for Microbiology. Susceptibility of the isolates to colistin was determined using Etest strips following the manufacturer’s guidelines (AB bioMérieux, Basingstoke, United Kingdom) and the guidelines of the Antibiogram Committee of the French Society for Microbiology. The strains with MIC < 2 mg/l were defined as susceptible to colistin for P. aeruginosa  and A. baumannii .

In nine patients of the multidrug-resistant strain group, computed tomography of the whole lung was obtained before and at the end of antimicrobial treatment according to the demand of the physician in charge. Contiguous axial 5-mm thick computed tomography sections of the whole lung were acquired at end-expiration before and after nebulized colistin. Volumes of gas and tissue and total lung volumes were computed.18,28Antibiotic-induced lung reaeration and decrease in lung inflammation after administration of nebulized colistin was measured as the increase in gas volume and decrease in tissue volume in lung regions characterized by multiple and disseminated foci of pneumonia and in lung areas characterized by confluent bronchopneumonia.19 

The trial was designed to demonstrate the efficacy of nebulized colistin for treating VAP caused by multidrug-resistant P. aeruginosa  or A. baumannii . The primary efficacy analysis assessed the noninferiority of clinical cure rate between multidrug-resistant strain and sensitive strain groups. Noninferiority of nebulized colistin was demonstrated if the lower limit of the one-sided 95% CI for difference in clinical cure rate was more than −16%. The selection of the noninferiority margin was determined by combining statistical reasoning and clinical judgment. The average clinical cure rates of VAP caused by sensitive P. aeruginosa  and A. baumannii  in our ICU in the past year were 60 and 85%, respectively. Considering that, in the study population, we would include 75% of patients with VAP caused by P. aeruginosa  and 25% of patients with VAP caused by A. baumannii , the mean clinical cure rate of VAP was estimated to be 66%. Because the multidrug-resistant strain group was a different population with longer length of mechanical ventilation and more episodes of VAP before inclusion, we did consider that 50% of clinical cure rate in these patients was clinically relevant. The rate of 50% is in accordance with the lower limit of cure rate reported in the literature in patients with VAP caused by sensitive P. aeruginosa  and A. baumannii  treated with appropriate intravenous antibiotics.29–32As a result, the noninferiority margin of 16% was determined (66–50% = 16%). Based on the inclusion ratio of 1:3 (one patient included in the multidrug-resistant strain group for three patients in the sensitive strain group), 32 and 96 patients were needed, respectively, in each group.

Risk for developing antibiotic resistance after administration of nebulized colistin in patients with recurrent pneumonia, assessment of serum colistin concentrations, determination of colistin-induced changes of computed tomography gas and tissue volumes, and assessment of antibiotic-induced nephrotoxicity were statistically analyzed.

Categorical variables were expressed as percentages and continuous variables as mean ± SD or as median and 25%, 75% interquartile range. A two-tailed hypothesis was tested in the statistical methods. Differences between sensitive strain and multidrug-resistant strain groups concerning clinical characteristics, duration of mechanical ventilation, length of stay and percentages of persisting VAP, recurrent VAP, and superinfection were compared using the chi-square test, bilateral unpaired Student t  test, or Mann–Whitney rank-sum test according to the data distributions. Computed tomography changes in gas and tissue volumes before and after nebulized colistin therapy were compared using bilateral paired Student t  test. Two-way analysis of variance for repeated measures was used to compare changes in CPIS and serum creatinine between groups. Statistical analysis was performed using SPSS 13.0 and SigmaStat 2.03 (SPSS, San Rafael, CA). A P  value of less than or equal to 0.5 was considered statistically significant.

Forty-three patients were prospectively included in the multidrug-resistant strain group. Ten patients had VAP caused by P. aeruginosa  or A. baumannii  susceptible only to colistin; 33 patients were infected by P. aeruginosa  or A. baumannii  resistant to all β-lactams and susceptible to colistin and aminoglycosides and/or ciprofloxacin. All patients had received inappropriate initial antimicrobial therapy. Twenty-eight patients were treated with nebulized colistin monotherapy. Fifteen patients were treated with nebulized colistin combined with a 3-day administration of intravenous aminoglycosides. The average duration of nebulized colistin administration was 12 days (range 7–19). In the sensitive strain group, 122 patients had VAP caused by P. aeruginosa  or A. baumannii  susceptible to at least one β-lactam. Eighty-four percent of patients had received initial appropriate empirical antibiotics. They were treated with intravenous β-lactam for 14 days combined with a 3-day administration of intravenous aminoglycoside or quinolone (fig. 1).

As shown in table 1, clinical characteristics of both groups were similar at ICU admission. At inclusion, VAP caused by P. aeruginosa  was more frequent in the sensitive strain group than in the multidrug-resistant strain group. In the multidrug-resistant strain group, duration of mechanical ventilation and ICU stay before initiation of antibiotics were significantly longer. In addition, previous administration of antibiotics and tracheostomy was more frequent. One third of patients of the multidrug-resistant strain group had at least two episodes of VAP before inclusion.

Twenty-nine of the 43 patients (67%) treated with nebulized colistin were clinically cured at the end of treatment compared with 81 of the 122 patients (66%) treated with intravenous β-lactams (difference: −1%, lower limit of 95% CI for the difference between success rate: −12.6%).

Nineteen of the 28 patients treated with nebulized colistin monotherapy and 10 of the 15 patients treated with nebulized colistin combined to a 3-day administration of intravenous aminoglycoside were clinically cured at the end of treatment. Clinical cure rate was not different between the patients treated with nebulized colistin monotherapy and those treated with nebulized colistin associated with a 3-day administration of intravenous aminoglycoside (67 vs . 68%, P = 0.94).

CPIS decreased significantly in patients successfully treated with antibiotics in both groups (fig. 2). As shown in table 2, 91% of patients with VAP caused by multidrug-resistant A. baumannii  were cured by nebulized colistin. All patients with VAP caused by sensitive A. baumannii  were cured by intravenous β-lactams. Treatment failure with persisting VAP caused by P. aeruginosa  was not statistically different between groups (P = 0.122). Recurrence of P. aeruginosa  VAP and VAP caused by superinfection was similar in both groups. The duration of mechanical ventilation after inclusion was longer in patients of the multidrug-resistant strain group. All-cause ICU mortality was similar between groups.

In the 122 patients of the sensitive strain group, 62% of initial P. aeruginosa  isolates were susceptible to all antipseudomonal antibiotics, 22% were resistant to carbapenems, and 16% were resistant to piperacillin and/or ceftazidime. Among 32 patients with persisting and recurrent P. aeruginosa  VAP despite antibiotic treatment, the resistance of P. aeruginosa  increased significantly: six strains became resistant to all β-lactams (table 3).

In the 43 patients of the multidrug-resistant strain group, colistin MIC at inclusion was 1.5 (0.75–2) mg/l for P. aeruginosa  strains and 0.12 (0.09–0.19) mg/l for A. baumannii.  Among 16 patients with persisting or recurrent VAP caused by P. aeruginosa , four strains became susceptible to β-lactams. Two patients had been treated with nebulized colistin monotherapy and two by nebulized colistin combined to a 3-day administration of intravenous aminoglycosides. In one patient treated with nebulized colistin monotherapy, colistin MIC increased from 0.75 to 3 mg/l. In another patient treated with nebulized colistin combined with 3-day intravenous aminoglycoside, colistin MIC did not change after the first VAP recurrence but increased from 1.5 to 3 mg/ml during the second VAP recurrence, after an additional 14-day administration of nebulized colistin.

In seven patients successfully treated with nebulized colistin, gas volume increased and tissue volume decreased significantly in regions with confluent pneumonia (fig. 3). They remained unchanged in regions with normal lung and/or disseminated foci of pneumonia. In two patients in whom nebulized colistin failed to treat VAP, tissue volume increased in lung regions with normal lung and/or disseminated foci of pneumonia (+177 and +198 ml) and decreased in regions of confluent pneumonia (−73 and −65 ml). Illustrative examples are shown in fig. 4.

Colistin serum concentrations were measured in 16 patients. Peak colistin concentrations were not different between day 2 and day 3. Trough colistin concentrations were significantly higher at day 3 compared with day 2 (fig. 5) and were not different between patients treated with nebulized colistin monotherapy and patients treated with nebulized colistin combined to a 3-day administration of intravenous aminoglycoside.

Serum creatinine remained stable within the treatment period in both groups of patients (fig. 6). At the end of treatment, increase of serum creatinine more than 1.5 times its baseline value was found in 8% of patients treated with intravenous β-lactam combined with a 3-day intravenous administration of aminoglycoside or quinolone and 12% in patients treated with nebulized colistin (P = 0.47). Per-treatment changes in serum creatinine were not different between patients treated with nebulized colistin monotherapy and patients treated with nebulized colistin combined to a 3-day administration of intravenous aminoglycoside.

Main results of the study are nebulized colistin is effective to treat VAP caused by multidrug-resistant P. aeruginosa and A. baumannii ; the clinical cure rate is noninferior to that obtained in VAP caused by susceptible P. aeruginosa and A. baumannii ; the risk of developing colistin resistance after nebulization is low; and nebulized colistin does not increase the risk of kidney failure, although repeated nebulization induces systemic accumulation.

Most multidrug-resistant Gram-negative pathogens exhibit resistance to almost all antibiotics except colistin. Experimental studies provide evidence for high antibiotic lung deposition and rapid bacterial killing after nebulization of antibiotics.10,15,16Therefore, the rationale for treating VAP caused by multidrug-resistant pathogens by nebulized colistin is strong. Nebulized colistin either as adjunct to systemic antibiotics32–34or as monotherapy35,36was previously reported in patients with VAP or tracheobronchitis. In these retrospective studies, neither the dose nor the conditions of nebulization were optimized. Our prospective study is the first to report the effectiveness of high-dose nebulized colistin as a treatment of VAP caused by multidrug-resistant P. aeruginosa  or A. baumannii .

Patients with VAP caused by susceptible strains and treated with a combination of intravenous β-lactams and aminoglycosides served as a control group. An a priori  hypothesis of a noninferior clinical cure rate of nebulized colistin compared with conventional bitherapy was made and defined as more than −16% difference of cure rate. Our aim was to prove that nebulized colistin can kill multidrug strains and cure lung infection. Recurrence after initial cure of VAP was therefore not regarded as a failure of antimicrobial therapy, considering that persisting reservoirs are inaccessible to either intravenous or nebulized colistin. Anyway, late recurrence of VAP was not different between groups.

In the multidrug-resistant strain group, nebulized colistin was given either alone (n = 28) or in combination with a 3-day administration of intravenous aminoglycosides (n = 15), based on clinical decision of attending physician. Many experimental and clinical studies suggest a poor efficiency of intravenous aminoglycosides for treating VAP. After intravenous administration, aminoglycosides tissue concentrations equal to or below MIC have been reported in animals with normal or infected lungs.15,37–39Aminoglycosides bronchial concentrations equal to or below MIC have been reported in patients with cystic fibrosis or VAP.40–43Two meta-analysis have demonstrated the lack of superiority of β-lactam–aminoglycoside combination to β-lactam monotherapy for treating various causes of sepsis.44,45These meta-analyses however suffer from important limitations: heterogeneity of sepsis in which the comparison was made; randomized clinical trials performed more than 15 yr ago; and comparison of different β-lactams in the tested treatment arms. As a consequence, many physicians still believe that patients with VAP might benefit from β-lactam–aminoglycoside combination. The same clinical cure rate was obtained in the 28 patients treated with nebulized colistin monotherapy and in the 15 patients treated with nebulized colistin–aminoglycoside combination. This result suggests that clinical outcome of VAP could be attributed to nebulized colistin rather than intravenous aminoglycosides.

In the sensitive strain group, clinical cure rate was observed in two thirds of patients, as previously reported.18,31,46In the multidrug-resistant stain group, clinical cure rate was noninferior to that observed in the sensitive strain group. The clinical benefit was associated with radiological improvement as demonstrated in figures 3and 4. It has to be pointed out that patients infected by multidrug-resistant strains were more severe than patients infected by sensitive strains: they had stayed longer on mechanical ventilation before inclusion; 95% of them versus  71% had received multiple antibiotics before inclusion; 28% of them versus  10% had two or more previous episodes of VAP, attesting of their inability to activate sufficient host response against infections; and 100% of them versus  16% had received inappropriate initial empirical antimicrobial therapy, a condition known to be associated with increased mortality and morbidity.47Therefore, the equivalence in terms of clinical cure rate clearly suggests a strong efficacy of nebulized colistin.

Such a benefit was obtained at two conditions. First, factors influencing distal lung deposition were optimized during each nebulization.18,48Second, daily high-dose aerosols were administered, taking into consideration that colistin is both a concentration- and time-dependent antibiotic12,13,49and has a limited systemic diffusion even in presence of injury of the alveolar–capillary barrier.10As previously recommended, nebulization dose of antibiotics should have been calculated as intravenous dose plus extrapulmonary deposition.48According to this rationale and considering an intravenous daily dose of 40,000 IU/kg and an extrapulmonary deposition of 40%, a nebulized dose of 56,000 IU/kg of colistimethate should have been administered. Because of its low systemic diffusion and its bactericidal profile, it was decided to increase colistin dose to 200,000 IU/kg, and 5 million IU were nebulized every 8 h, a dosage two to three times higher than previously reported.32,34In patients with cystic fibrosis where the target is the bronchial tree, the nebulization of 2 million IU of colistin provides high bronchial concentrations.50In patients with VAP where the target is the alveolar space, much higher doses are required to obtain at least fivefold the MIC at the site of infection. Recently, a pharmacokinetic study performed in patients with ventilator-associated tracheobronchitis has shown that a dose of 1 million IU of nebulized colistin every 8 h was not adequate to treat lung infection caused by multidrug-resistant strains.51In the current study, no adverse respiratory and bronchial effects were observed after nebulization, confirming a recent experimental study.10 

Whether using nebulized or intravenous colistin for treating VAP caused by multidrug-resistant Gram-negative bacteria has been a controversial subject.32,33,46,52The current study brings convincing evidence that nebulized colistin is clinically efficient, whereas experimental and clinical studies report a poor lung penetration of intravenous colistin.10,11Ninety-one percent of patients with VAP caused by multidrug-resistant A. baumannii  were cured by nebulized colistin. This result is likely explained by high tissue concentration to MIC ratio obtained after administration of nebulized colistin. Only 59% of patients with VAP caused by multidrug-resistant P. aeruginosa  were cured by nebulized colistin likely because of a less favorable tissue concentration to MIC ratio. In a recent experimental study,10MIC of colistin for P. aeruginosa  was 2 μg/ml, whereas median peak tissue concentrations were 2.8 μg/g (tissue concentration to MIC ratio = 1.8). MIC of colistin for A. baumannii  is 0.2 μg/ml, and assuming a median colistin peak tissue concentration of 2.8 μg/g (tissue concentration to MIC ratio = 18). Colistin being a concentration-dependent antibiotic, it is easy to understand why A. baumannii  strains were massively eradicated after the end of treatment. VAP caused by P. aeruginosa  had a higher incidence of recurrence and superinfection possibly related to the too short duration of treatment resulting from extubation before the 2-week scheduled regimen, as previously reported.22 

In the sensitive strain group, 75% of patients whose VAP persisted or relapsed after intravenous antimicrobial therapy acquired resistance to β-lactams. Interestingly, in the multidrug-resistant group, 25% of patients whose VAP caused by P. aeruginosa  persisted or relapsed after nebulized colistin, recovered susceptibility to β-lactams. These findings are in accordance with two previous studies showing that nebulized antibiotic decreases bacterial resistance in patients with VAP.18,53 

It has been reported that prolonged use of intravenous colistin predisposes to VAP caused by pandrug-resistant P. aeruginosa , likely due to colistin poor lung tissue penetration and low concentrations at the site of infection.54,55In patients treated with nebulized antibiotics, much higher tissue concentrations are present in infected lung regions,10thereby preventing selection of resistant strains. In the current study, MIC of P. aeruginosa  strain to colistin increased in two patients after nebulized colistin administered during 14–21 days, indicating a low rate of acquisition of resistance.

Experimental data suggest a decreased colistin systemic exposure after 24 h of nebulization.10The current study shows that colistin trough plasma concentration significantly increased between day 2 and 3, suggesting colistin accumulation with time as a result of slow systemic passage through the alveolar–capillary membrane.56,57Renal function impairment was observed in 12% of patients treated with nebulized colistin, and evolution of serum creatinine during the treatment was similar in both groups. Therefore, high doses of nebulized colistin can be considered as safe when administered during 14 days. We recommend however to measure colistin serum concentration at day 7, particularly in patients with preexisting alterations of renal function.9,32 

In conclusion, nebulized colistin at high dose is effective and safe for treating VAP caused by multidrug-resistant P. aeruginosa and A. baumannii . It provides an attractive alternative in the face of the increasing incidence of VAP caused by multidrug-resistant Gram-negative pathogens in critically ill patients. Further randomized controlled studies comparing intravenous and nebulized colistin for treating VAP caused by multidrug-resistant P. aeruginosa and A. baumannii  are required.

Guang-Ju Zhou, M.D., Research Assistant, Mao Zhang, M.D., Ph.D., Professor of Emergency Medicine, Medical Director (Department of Emergency Medicine, Second Affiliated Hospital, Zhejiang University, School of Medicine, Hangzhou, China); Charlotte Arbelot M.D., Praticien Hospitalier, Hélène Brisson, M.D., Praticien Hospitalier, Antoine Monsel, M.D., Chef de Clinique Assistant, Ivan Goldstein, M.D., Ph.D., Research Assistant, Alfonso Sartorius, M.D., Research Fellow, Marine Lecorre, M.D., Research Fellow, Fabio Ferrari, M.D., Ph.D., Research Fellow, Corinne Vezinet, M.D., Praticien Hospitalier (Multidisciplinary Intensive Care Unit, Department of Anesthesiology and Critical Care Medicine, La Pitié-Salpêtrière Hospital, Assistance-Publique-Hôpitaux-de-Paris, UPMC Univ Paris 06, France); Claudia Gutierrez, M.D., Research Fellow, Department of Anesthesiology, Faculty of Medicine Federal University from Rio Grande do Sul, Hospital das Clinicas de Porto Alegre, Brazil; Charles-Hugo Marquette, M.D.,Ph.D., Professor of Pneumology, Medical Director of the Department of Respiratory Diseases, Pasteur hospital, University of Nice-Sophia Antipolis, Nice, France; Olivier Petit-Jean, Ph.D., Professor of Pharmacology, Director of the Department of Pharmacology, Avicenne Hospital, Assistance Publique-Hôpitaux de Paris, Bobigny, France; Christine Bernard, Pharm D, Praticien Hospitalier, Jérôme Robert, M.D., Ph.D., Professor of Bacteriology (Department of Bacteriology, La Pitié-Salpêtrière Hospital, Assistance-Publique-Hôpitaux-de-Paris, UPMC Univ Paris 06), Christian Funck-Brentano, M.D., Ph.D., Professor of Pharmacology, Medical Director, Department of Pharmacology, La Pitié-Salpêtrière Hospital, UPMC Univ Paris 06, France.