The growing epidemic of multidrug-resistant (MDR) gram-negative bacteria has led clinicians to prescribe colistin. However, due to inadequate penetration in the lung parenchyma, the effectiveness of intravenous (IV) colistin therapy for pneumonia has been questioned.1

Although inhaled colistin has been used successfully to prevent and cure pulmonary infections in patients with cystic fibrosis colonized with Pseudomonas aeruginosa,2 there is limited data supporting effectiveness and tolerability of inhaled colistin therapy in non-cystic fibrosis patients with pneumonia due to MDR A. baumannii.3 We report here our experience with critically ill patients who received colistin inhalation therapy for the treatment of colistin-only-susceptible (COS) A. baumannii pneumonia.

During the study period, 12 (7 males) patients were treated with inhaled colistin for COS A. baumannii pneumonia. Table 1 describes the demographic and clinical features of these patients. The median age of the patients was 75 years, and median APACHE II score on the day of inhaled colistin treatment was 19.5 (range 12-29). Polymicrobial infections were observed in 5 patients. In 2 of these 5 patients (Case 9 and 11), P. aeruginosa was isolated from the same culture specimen. Methicillin resistant Staphylococcus aureus was isolated in the other 3 patients (Case 4, 8, 12). All patients were spontaneously breathing without mechanical ventilation. However, most patients were severely ill (10 were in intensive care unit), and had many underlying diseases. One patient (Case 3) had concomitant COS A. baumannii bacteremia. She received IV colistin for 3 days, and then switched to colistin inhalation with IV cefoperazone/sulbactam therapy. Because inhaled colistin was administered as definitive therapy in all patients, inhalation treatment started 3-5 days after symptom onset. Therefore, all patients had received other broad spectrum antibiotics before colistin inhalation therapy. Inhaled colistin monotherapy was administered to 4 patients (Case 1, 2, 5, and 7). In 5 patients, an additional antimicrobial agent with activity against gram-negative bacteria was given during the course of inhaled colistin treatment.

Favorable clinical response (cure or improvement) was observed in 10 of the 12 patients (Table 2). Follow-up cultures were available for all patients and A. baumannii was eradicated in 6 patients. The all-cause in-hospital mortality was 33.3% (4/12 patients). Pneumonia-associated mortality was 8.3% (1/12 patients). One patient died from subsequent clinical sepsis related to severe Clostridium difficile infection, by 47 days after resolving from pneumonia episode (Case 9). In addition, other 2 patients died from acute bleeding and myocardial infarction. Since A. baumannii isolates in our case series were resistant to other antimicrobials, we expected that the clinical outcomes could be attributed to nebulized colistin rather than systemic antimicrobial agents. However, the outcomes were different from our expectations. Among the 4 patients treated with colistin inhalation therapy alone, favorable clinical response was observed in 2 (50%) patients. Two of the 4 patients died during the hospitalization and one patient died from pneumonia. By contrast, all the 8 patients who received concurrent systemic antimicrobial agents showed favorable outcome, and only one died from acute myocardial infarction.

Colistin inhalation therapy was well tolerated (Table 3). During treatment, no patient experienced bronchospasm or neurotoxicity. Only one patient experienced fever and skin rash after colistin inhalation therapy (Case 7). The physician suspected colistin hypersensitivity and changed her antibiotic regimen. Two patients experienced Clostridium difficile associated diarrhea. Three patients suffered from acute kidney injury: 2 RIFLE-R, and one RIFLE-I. Among them, 2 patient's peak serum creatinine levels were within normal reference range (Case 4 and 12) and renal functions returned to baseline at the time of hospital discharge. One patient (Case 5) died and follow up of the creatinine level was not possible for this patient. None of the patients discontinued colistin treatment or required renal replacement therapy because of colistin induced nephrotoxicity. Colistin resistance in subsequent sputum culture isolates developed in 3 patients. Median interval between colistin inhalation therapy and resistance was 7 days (range 5-19 days). One patient (Case 4) received colistin inhalation therapy based on the result of culture study performed 5 days prior to the treatment. Subsequent sputum culture obtained on the treatment initiation day was reported as a colistin resistant A. baumannii thereafter. Despite colistin resistance, colistin inhalation therapy was continued and clinical outcomes were favorable in all of 4 patients. Burkholderia cepacia superinfection was observed in one patient (Case 8).

Colistin has a narrow spectrum of use and is primarily used for infections with P. aeruginosa and A. baumannii. Gram-positive bacteria, Burkholderia cepacia, Serratia marcescens, Moraxella catarrhalis, Proteus spp., Providencia spp., and Morganella morganii are inherently resistant to colistin.7 Nephrotoxicity and neurotoxicity are the main adverse effects of colistin treatment. Despite these shortcomings, IV colistin has been used as salvage therapy in the treatment of pneumonia caused by MDR or COS gram negative pathogens due to the absence of effective alternative options.8

There are several studies evaluating the use of nebulized colistin either as monotherapy or as combination with systemic antibiotics for the treatment of pneumonia associated with MDR pathogens (Table 4). In some case-series studies, colistin inhalation adjunctive to IV colistin therapy has shown high favorable clinical and microbiological responses.9-12 However, the results from studies including control groups are not consistent. Korbila, et al.13 demonstrated that the outcome of ventilator-associated pneumonia was better in patients who received colistin inhalation with IV colistin than those who received IV colistin alone. By contrast, other studies showed that addition of aerosolized colistin to IV colistin resulted in no additional benefit.5,14 Aerosolized colistin inhalation with systemic antibiotics other than colistin also showed favorable clinical (61-100%) and microbiologic responses (60.9-83.3%).3,15-22 These studies include different combinations of concurrent antimicrobial regimen; therefore, the optimal combination and dosage should be determined by further studies. Colistin inhalation monotherapy without concurrent IV antimicrobial is rarely reported. In a case series by Falagas, et al.23 two children received inhaled colistin monotherapy for tracheobronchitis and both cases showed a favorable response.

In our study, 8 patients who were treated with concurrent systemic antibiotics other than colistin, showed 100% clinical response, which was similar to the results from previous studies.3,15-22 However, colistin inhalation as monotherapy in our study showed only a 50% clinical response.

Four patients were treated with an inhaled colistin only, and the other 3 patients received concomitant IV antimicrobial, which has no gram negative activity. Understanding why physicians treated their patients with a colistin inhalation monotherapy is of interest. First, all of these patients were older and had neurovascular diseases. Accordingly, the physicians might be worried about neurotoxicity associated with systemic administration of colistin. The second reason is possible concern for nephrotoxicity. All the patients had hypoalbuminemia (2.4-2.8 g/dL) and two patients (Case 4 and 12) received IV vancomycin concurrently. The other 2 patients (Case 5 and 7) showed impaired renal function at that time. The third reason is likely financial. Except for co-infection by COS A. baumannii and colistin-resistant bacteria, The Korean National Health Insurance system often refuses reimbursement for colistin combination therapy with other antimicrobial agent having gram negative activity.

Two major risks are arising from the wide use of colistin: the emergence of colistin resistance and an increase of infections due to gram-positive and gram-negative bacteria that are inherently resistant to colistin. With respect to resistance, colistin heteroresistance and colistin resistance in A. baumannii have been reported in several studies.24-33 In Korea, colistin resistance in A. baumannii was reported as 9.1% and 30.6%, respectively, in two different reports.32,33 In our hospital, colistin resistance rate in A. baumannii has been less than 1%. Although some studies suggested that nebulized antibiotic decreases bacterial resistance,34,35 colistin inhalation monotherapy may be problematic for treatment of pneumonia caused by colistin heteroresistant A. baumannii. To prevent the appearance of colistin resistance and to improve treatment efficacy, the optimization of the colistin dose is essential. Although previous studies with inhaled colistin therapy did not report any colistin-resistant isolates in patients with persistent isolation of A. baumannii or P. aeruginosa, our study results support this concern in that 3 patients showed resistance to colistin in subsequent sputum culture isolates within 20 days of colistin inhalation.

The side effects of nebulized colistin include bronchoconstriction, cough, chest tightness, and apnea due to neuromuscular blockade.36 In this report, no patients were seen developing such adverse events. Our experience suggests that aerosolized colistin for pneumonia due to COS A. baumannii may offer advantages over systemic administration in respect to the occurrence of side effects. Theoretically, the use of the aerosolized colistin can minimize potential renal and neurologic toxicities because of negligible systemic absorption. A pharmacokinetic study of aerosolized colistin in cystic fibrosis patients revealed that the maximum observed serum colistin concentration was 0.178 mg/L, <10% of the level in sputum, and much less than the maximal level of 2.8-13.9 mg/L achieved by the IV route and this was undetectable 12 hours after the inhalation.2 We did not measure the plasma level of colistin in our patients, but it is reasonable to assume low-serum levels and therefore a lower risk of systemic side effects in our patients. In our cases, nephrotoxicity was higher than previous studies. This seems to be due to the severity of patients' diseases and concurrently administered glycopeptides. In addition, our definition of nephrotoxicity was more sensitive than in previous studies. Although the serum concentration is low, the possibility of systemic accumulation by repeated nebulization cannot be excluded. This could result in an increased risk for the selection of colistin resistant strains of A. baumannii.

Our study has some limitations. First, it was adapted from a single center, and a small case series were enrolled using a retrospective design. In addition, there we had no control group to compare our results. Second, neurotoxicity may not have been detected in our patients because most of them had underlying neurovascular diseases. Third, we identified Acinetobacter species by phenotypic method only; therefore, it is possible that non-baumannii species were included. Fourth, there is the possibility of isolated A. baumannii was not the etiology of pneumonia but colonization.

In summary, it is difficult to decide whether to use inhaled colistin monotherapy in pneumonia caused by COS A. baumannii. Moreover, superinfection with inherently colistin-resistant pathogens and the development of colistin resistant P. aeruginosa or A. baumannii should be prospectively monitored. The appropriate dose of aerosolized colistin, effective combination with other antimicrobial agents, and adequate indications remains to be determined in order to minimize these risks. A larger prospective and controlled study is needed.