Intravenous colistin methanesulfonate (CMS) use is limited by a low safety profile due to renal and neurotoxicity, particularly when high dosages are required to achieve pharmacokinetic/pharmacodynamic (PK/PD) targets, as in the case of nosocomial pneumonia that require sufficient lung penetration ^[1][2][28,29]. In fact, clinical and pharmacologic data reveal that IV CMS demonstrates limited efficacy against respiratory tract infections ^[3][30]. In contrast, nebulized CMS attains high lung tissue concentrations ^[4][31], enabling rapid bacterial killing ^[5][32]. Optimized nebulization techniques, which make use of specially designed ventilation circuits and appropriately adjusted respiratory settings, result in the desired lung tissue penetration, contributing to efficient pathogen eradication, as demonstrated in an animal study ^[6][33]. By targeting the delivery of colistin, nebulized administration increases antibiotic concentrations in the epithelial lining fluid (ELF), especially in high doses and regardless of concomitant IV administration ^[7][19]. Intravenous dosing recommendations may be inadequate for critically ill patients with HAP/ventilator-associated pneumonia (VAP), mainly due to altered pharmacokinetic profiles and decreased ELF penetration ^[2][29]. This targeted delivery not only boosts treatment efficacy but also reduces systemic exposure and potential adverse events, such as nephrotoxicity. High-dose drug nebulization, such as 5 MIU of CMS every 8 h, allows for high tissue concentrations while maintaining plasma levels below nephrotoxic levels ^[7][19]. Intrapulmonary CMS partially diffuses into the systemic compartment, with only 17% of the nebulized dose being rapidly eliminated by the kidneys ^[8][34]. It becomes evident that this approach may contribute to reducing healthcare costs due to prolonged and complicated hospitalization, as well as reducing renal replacement therapy interventions and improving patient therapy tolerance overall.

Nebulized CMS can also be safely administered at high and very high doses ^[9][35], owing to its favorable PK/PD properties and excellent bronchial tolerability, as evidenced by its use as a monotherapy in patients without cystic fibrosis ^[10][36]. It is a concentration-dependent antibiotic with a post-antibiotic effect. High-dose nebulization can be especially beneficial for critically ill patients with VAP due to the presence of a high bronchial inoculum ^[8][34]. The infected compartments of the lung act as a CMS reservoir, where it undergoes slow hydrolysis into colistin, providing continuous bacterial killing. Therefore, high colistin concentrations are achieved in the affected lung regions, enhancing the effectiveness of antibiotic treatment ^[11][37]. Another crucial consideration when choosing a high-dose regimen is the potential emergence of hetero-resistance due to bacterial exposure to suboptimal colistin dosages ^[12][38]. Overall, the localized distribution of nebulized CMS is key to unlocking its potential in terms of both efficacy and safety. To maintain colistin’s efficacy and prevent the development of resistance, it is essential to implement optimized dosing and administration techniques for high-dose CMS.

The optimal high dose of nebulized CMS has not currently been defined. However, the effectiveness and safety of high doses of up to 15 MIU aerosolized CMS per day have been evaluated ^[13][14][15][20,21,22]. The CMS pharmacokinetic and pharmacodynamic properties, the excellent bronchial tolerability ^[15][22], and the low systemic absorption and associated toxicity ^[13][20] suggest that doses as high as 15 MIU can be recommended for difficult-to-treat Gram-negative lower respiratory tract infections. The administration of high-dose aerosolized CMS results in the achievement of elevated lung tissue concentrations if the nebulization procedure is optimized ^[7][19], and it compensates for colistin loss due to extrapulmonary disposition. In addition, the inoculum effect of CMS ^[8][34] and its slow hydrolysis to colistin in the lung support the administration of high doses. Minimal systemic exposure of colistin after nebulization of high doses of CMS and slow hydrolysis of CMS into colistin in plasma result in low serum concentrations, regardless of nebulization dose ^[13][20].

According to CMS product characteristics, 1 MIU of CMS should be dissolved in 3 mL of normal saline solution; thus, a 15 mL solution is required to deliver a dose of 5 MIU. Since the volume of most nebulization chambers ranges between 6 and 10 mL, the nebulizer needs to be filled at least twice. This increases the nurse’s workload, prolongs nebulization time beyond 60 min, and carries the risk of incomplete drug administration. A recent study ^[16][39] demonstrated that a reduction of diluent volume to 6 mL for nebulization of dilution of 4 million IU resulted in shorter nebulization time. In addition, compared with the 12 mL solution, the stability of the 6 mL solution was increased, and the nebulization time was significantly shortened. Moreover, no modification was observed in aerosol particle characteristics and plasma and urine pharmacokinetics. Therefore, dilution of 5 MIU of CMS with less than 10 mL of normal saline is possible.

CMS and colistin are not stable in aqueous media ^[16][17][39,40]; thus, reconstitution should be performed just before nebulization. Fatal acute respiratory distress syndrome has been described in a patient with cystic fibrosis and chronic airway infection with Pseudomonas aeruginosa after administration of a premixed CMS nebulization that was 5 weeks old ^[18][41]. In this patient, conversion of the colistin prodrug to the biologically active form of colistin due to prolonged storage of the aqueous solution was considered the likely cause of death.

The deposition of aerosol in the infected lung is directly influenced by the size of aerosolized particles and the type of nebulizer. Particles > 5 μm tend to deposit in the ventilator circuit and large airways, and the optimal median mass aerodynamic diameter ranges between 0.5 and 3 μm. Nebulization of high-dose colistin can be performed via jet (JN), ultrasonic (USN), and vibrating mesh nebulizers (VMN) which can generate aerosol particles with a diameter of < 5 μm.

JN are the most commonly used devices in invasively ventilated patients ^[19][42]. Nebulization is performed with exposure of the antibiotic to a highly pressurized air or oxygen flow delivered intermittently or continuously. Delivery of the drugs is not constant. Advanced ventilators perform synchronized nebulization, in which a fraction of the inspiratory flow is used to power nebulization. Compared with USN or VMN, JN has been shown to have the lowest efficiency due to high residual volume, accumulation of the drug in the circuit, and loss of drug through the expiratory limb ^[19][20][42,43]. According to previous studies, the best position for drug delivery is in the inspiratory limb 15 cm from the ventilator ^[21][22][44,45]. In recent studies, JN has been used for the administration of high-dose nebulized colistin ^[23][24][25][46,47,48].

In USN, a piezoelectric quartz crystal generates aerosol through vibration ^[19][20][42,43]. The efficiency of drug delivery is better than JN. In a previous study comparing the efficiency of JN and US in intubated patients, it was demonstrated that pulmonary deposition as a percentage of initial nebulizer activity was significantly greater with USN ^[26][49]. The best position for optimized drug delivery appears to be on the inspiratory limb at approximately 15 cm from the Y piece ^[21][44]. The main disadvantage of USN is the increase in solution temperature by 10 to 15 °C after 5 min of nebulization, which may affect the stability of nebulized CMS ^[19][20][42,43]. However, limited data exist regarding the administration of high-dose nebulized colistin with USN ^[27][50].

VMN consists of a small drug reservoir placed above a dome-shaped aperture plate with more than 1000 funnel-shaped apertures attached to a piezoceramic element ^[19][20][42,43]. The vibration of the aperture plate pumps liquid through the apertures, where it is broken into fine particles between 3 and 5 μm in size. The superiority of VMN over JN for nebulized antibiotics has been demonstrated in both in vitro and in vivo studies ^[9][28][35,51], and nebulization with VMN is recommended to optimize drug delivery to the lung. VMN is more efficient than JN and USN due to lower residual drug volume ^[21][44], and it does not affect solution temperature and, thus, stability ^[19][20][42,43]. In invasively ventilated patients, continuous rather than inspiration-synchronized nebulization is preferable as the latter requires extensively prolonged nebulization time ^[9][35]. Continuous nebulization results in the delivery of highly concentrated aerosol to the lung due to the bolus effect ^[9][35]. In addition, VMN operation without an external gas source and maintenance of ventilation delivery parameters are major advantages for clinicians ^[20][43]. Administration of high-dose nebulized colistin is mainly performed by VMN, as observed in recent studies ^{[7][13][14][15][29][30]}[19,20,21,22,52,53]. In order to optimize drug delivery, the best VMN position in the circuit is 15 cm from the Y-piece in the inspiratory limb ^[21][22][31][44,45,54]. A comparison of the nebulizer generator devices is presented in Table 1 ^{[9][19][20][21][28][32]}[13,35,42,43,44,51].

Limiting inspiratory flow velocity is important because it reduces turbulence, aerosol impaction, and drug deposition in the circuits, thereby optimizing lung deposition. To achieve that goal, it is important to use specifically designed smooth angles and inner surface circuits ^[9][20][35,43]. In addition, a specific ventilator setting should be applied during nebulization: volume-controlled ventilation with constant inspiratory flow, inspiratory/expiratory ratio < 50%, tidal volume at 8 mL/kg, respiratory frequency 12–15 bpm, and minimum bias flow (2 L/min) ^[9][20][35,43]. A plateau end-inspiratory pause of 20% of the duty cycle and positive end-expiratory pressure of 5 to 10 cm H₂O should be applied to promote alveolar deposition ^[9][35]. During nebulization, the administration of a short-acting sedative or the transient increase in sedation is important to avoid patient–ventilator desynchrony ^[9][20][35,43]. In addition, the heat and moisture exchanger should be removed and heated humidification interrupted to avoid hygroscopic growth and massive trapping of aerosolized particles ^[9][20][35,43]. The placement of a filter on the expiratory limb is necessary to protect the ventilator flow device, and a filter change should be performed after each nebulization to avoid obstruction.

There is significant uncertainty relating to the benefits, optimal dosage, and clinical efficacy of nebulized colistin, mainly due to a lack of accurate PK data. The scarcity of these data may be explained, in part, by the complexity of PK and physicochemical properties of the drug, an intravenous form of which has been shown to exhibit inadequate penetration into lung tissue ^[2][33][29,55]. Existing PK data of nebulized colistin in critically ill patients with VAP caused by resistant pathogens were derived, so far, from a handful of small cohort studies.

A first effort to describe colistin concentrations in ELF of 20 critically ill patients with ventilator-associated tracheobronchitis (VAT) caused by polymyxin-only susceptible GNB was conducted by Athanassa et al. ^[4][31] via a high-pressure liquid chromatography (HPLC)-based method, which provides a more accurate analysis of colistin and its prodrug CMS and concentrations than the previously used microbiological assays ^[34][56]. A dose of 1 MIU, dissolved in 3 mL of half-normal saline, was administrated for over 30 min every 8 h via a vibrating-mesh nebulizer, the theoretically standard of care device, which generates the preferential particle size of the drug, achieving high concentrations in lung parenchyma ^[35][57]. After performing mini bronchoalveolar lavage, authors evaluated formed colistin ELF concentrations. Median values were 6.7 (4.8–10.1), 3.9 (2.5–6.0), and 2.0 (1.0–3.8) mg/L at 1, 4, and 8 h, respectively, and fivefold higher than those in plasma. However, measured ELF concentrations of the drug were below the MIC of isolated pathogens 4 h after inhalation, indicating the sub-optional dosage of 1 MIU of nebulized CMS.

A more proper characterization of the nebulized colistin PK was reported in the study of Boisson et al. ^[36][27], who investigated ELF and plasma CMS and formed colistin concentrations in 12 patients with VAP after inhalation of 2 MIU of CMS dissolved in 10 mL of saline and nebulized for over 30 min, followed 8 h later by the same dose of IV CMS. According to the PK model applied in this researchtudy, the range of colistin concentrations after inhalation was 9.53 to 1137 mg/L in ELF, and 0.15 to 0.73 mg/L in plasma, indicating that measured CMS and colistin concentrations in ELF were 100- to 1000-fold higher than those in plasma. Intravenous administration did not achieve therapeutic levels in the infection site ^[36][27].

The study by Gkoufa et al. was the first to evaluate the PK of CMS and formed colistin in ELF after high doses of nebulized CMS (3 MIU and 5 MIU) ^[7][19]. In detail, the study population included 30 patients with VAP divided into three equal groups. Ten patients received concomitantly IV and nebulized CMS of 3 MIU administered in 30 min via a vibrating-mesh nebulizer, 10 patients received 3 MIU of nebulized CMS as monotherapy, and for another 10 patients, 5 MIU of nebulized CMS was administrated as monotherapy. The applied PK model predicted the concentrations of CMS and formed colistin in ELF over 24 h, as well as estimated the unbound fraction of formed colistin. After a dose of 3 and 5 MIU of CMS, predicted trough concentrations of formed colistin in ELF were 120.4 mg/L and 200.7 mg/L, respectively. These concentrations ranged from more than 100- to 600-fold higher than those in plasma and more than 100-fold higher than the median MIC (i.e., 1 mg/L) of isolated pathogens. Regarding the free ELF concentrations of formed colistin, values were approximately 1- to 10-fold higher than the median MIC (i.e., 1 mg/L); however, after evaluating IV CMS administration as monotherapy, the formed colistin concentration in ELF was predicted to be much lower (>10-fold) compared with the nebulized groups. Moreover, IV administration of CMS did not contribute significantly to ELF-formed colistin concentrations. No safety issues of higher doses of nebulized CMS were raised from this researchtudy.

A more recent PK study using a high dose of nebulized CMS (5 MIU) investigated ELF CMS and colistin concentrations in seven critically ill patients with VAP ^[15][22]. PK results reported that one hour after nebulization, the median colistin and CMS concentrations in ELF were 121.7 (40.1–143.1) mg/L and 1445.3 (236.2–1918.2) mg/L, respectively, while twelve hours after nebulization, the median colistin and CMS concentrations were 122.6 (43.3–130) mg/L and 522.3 (222.3–636.5) mg/L, respectively. Colistin concentrations were far above the median MIC (1 mg/L) of isolated Acinetobacter baumannii, a finding of great importance not only in effectively treating lung infections but also in preventing the risk of acquisition of colistin resistance.

Systemic PK of high doses of nebulized CMS was assessed by Benitez-Cano et al., who investigated plasma colistin concentrations in 27 patients with VAP or HAP, 15 receiving 3 MIU and 12 receiving 5 MIU of CMS, dissolved in 6 mL and in 10 mL of saline, respectively, and nebulized for 30 min ^[13][20]. In this researchtudy, two types of nebulizers were used, a vibrating-mesh nebulizer in 17 patients and a jet nebulizer in 10 patients. Median (IQR) quantifiable formed colistin concentrations in plasma at 1, 4, and 8 h after nebulization of 3 MIU and 5 MIU of CMS were below 0.20 mg/L and 0.24 mg/L, respectively. Even high doses of nebulized CMS were proved to achieve undetectable or very low plasma colistin concentrations (<1 mg/L), being at the same time safe and well-tolerated, as the reported concentrations were much lower than those potentially reported to cause nephrotoxicity (~2.5 mg/L) ^[37][58]. A major limitation of the study was the lack of intrapulmonary PK data on nebulized colistin.

Similarly, a previously reported study evaluated the PK of CMS and formed colistin in plasma, besides those in ELF, after high doses of nebulized CMS (3 MIU and 5 MIU) using a population PK approach ^[7][19]. Authors indicated that free plasma concentrations of formed colistin after nebulization were minimal and below 1 mg/L across 24 h for both dosing groups and lower than those defined to cause nephrotoxicity, a result consistent with previous studies ^[13][36][20,27].

As mentioned above, the available nebulization devices present considerable differences regarding their way of function and particle generation, and their availability mainly guides the decision for type selection. Moreover, after reviewing the literature for studies using nebulized colistin, researchers will discover a knowledge gap regarding the preferred nebulizer device, which may achieve sufficiently small particles of the drug in order to reach the pulmonary alveoli. Although published data support the superiority of vibrating mesh over jet and ultrasonic nebulizers, mainly due to advantages related to their manufacturing characteristics, lung PK data comparing different devices are missing, while in the study of Benitez-Cano et al. plasma colistin concentrations were higher with the use of vibrating-mesh compared to jet nebulizers ^[13][20]. A recent study by Kyriakoudi et al. compared ELF and plasma PK data of patients with VAP receiving 2 MIU of CMS either via a vibrating mesh or a jet nebulizer ^[38][59]. The maximum colistin concentrations in ELF, obtained with a vibrating mesh nebulizer were 10.4 (4.7–22.6) mg/L, while maximum ELF colistin values obtained with a jet nebulizer were 7.4 (6.2–10.3) mg/L. Regarding the C_max and C_min plasma formed colistin concentrations for the VMN were 2.6 (2.0–3.5) mg/L and 0.2 (0.1–0.3) mg/L, respectively, whereas for the JN, 0.3 (0.3–1.6) and 0.1 (0.1–0.2) md/L accordingly. Thus, the authors concluded that both nebulizers led to comparable formed colistin concentrations in ELF, providing a valuable finding in the field, as clinicians could probably have a safe and reliable alternative, considering the availability of devices and consumables in every hospital.

The aforementioned data indicate that higher doses of nebulized colistin may achieve adequate concentrations in lung compartments and, importantly, well above the MICs of isolated pathogens while, at the same time, eliminating systemic exposure and risk of nephrotoxicity and overcoming the obstacles of low penetration of intravenous colistin in ELF and adsorption of the drug to surfaces of sampling devices at low concentrations. However, looking across studies, plasma concentrations of formed colistin present variability after nebulization of different doses of CMS and do not follow a linear increase after administration of higher doses. Studies using low doses of nebulized CMS reported higher concentrations of colistin in plasma ^[4][38][31,59] compared with studies administrating high doses of nebulized CMS, which demonstrated low plasma colistin concentrations ^[7][13][19,20]. These discrepancies may be explained either by the PK properties of the drug or by the potential hydrolysis of CMS during analytic procedures, facts that may have biased measured colistin concentrations in plasma ^[39][60]. Notably, evaluating lung interstitial colistin concentrations in patients with VAP remains a challenge that requires a thorough investigation and approach and is still not fully elucidated. Factors that mainly hampered the achievement of therapeutic levels of this still potentially valuable antibiotic in the lung, and may influence measured concentrations, probably include the nebulized colistin dose and dosage interval, the possible contamination of the bronchoscope by bronchial secretions during the bronchoalveolar lavage, the adsorption of colistin to plastic surfaces, the methods used for the determination of colistin concentrations, and the population study—critically ill patients have different kinetics from other patients ^[40][61]. PK data of nebulized colistin should be further accompanied by observational studies based on clinical efficacy in order to reinforce its use by ensuring the beneficiary effect in infection-related outcomes.