INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: aerosol; amikacin; piglets; mechanical ventilation

Ventilator-associated pneumonia is the most frequent nosocomial infection in the critically ill, and gram-negative bacilli are the causative microorganisms in more than 60% of cases (1, 2). Because of the severity of these infections and the increasing incidence of multiresistant causative organisms, intravenous aminoglycosides are commonly used in association with betalactam antibiotics. The tissue concentration at the site of infection is the main determinant of bactericidal efficiency, whereas toxicity depends on the trough plasma concentration (3, 4). Intravenous aminoglycosides easily penetrate in lung parenchyma and bronchial secretions according to the plasma-tissue concentration ratio. However, lung tissue concentrations remain small (5) because plasma levels are kept low deliberately to avoid toxicity. Aerosol administration offers the theoretical advantage of high concentrations of antibiotics at the site of infection together with a low systemic absorption resulting in reduced renal toxicity. We previously demonstrated that the intratracheal administration of colistin could significantly reduce the incidence of ventilator-associated pneumonia in a large series of critically ill patients (10). In fact, little information exists on the deposition of aerosolized antibiotics into the pulmonary parenchyma during mechanical ventilation. A number of human studies have measured concentrations of inhaled aminoglycosides in bronchial secretions (11) and have demonstrated that these concentrations are more indicative of the deposition of the antibiotic in the central airways than in the alveolar compartment (16). Because of methodological limitations regarding tissue sampling in humans, an animal model was required to assess the quantitative deposition of inhaled aminoglycosides in the whole lung parenchyma. According to previous studies that identified parameters required for an optimal delivery of aerosols, an ultrasonic nebulizer equipped with a large reservoir was selected to administer amikacin in a previously validated experimental model of mechanically ventilated piglets (17, 18).

The aim of the study was to compare the lung tissue concentrations, the regional distribution, and the plasma pharmacokinetics of amikacin administered either intravenously or by aerosol in two groups of mechanically ventilated piglets with healthy lungs. Additionally, retention in the circuit and the exhaled fraction were measured to determine the pulmonary availability of amikacin. In the group of animals that received nebulized amikacin, fractionated urinary output was performed to evaluate the amount of amikacin that reached the systemic compartment after lung deposition (systemic bioavailability).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Preparation and Aerosol Generation

Thirty-six healthy bred domestic Largewhite-Landrace piglets (20 ± 2 kg) were anesthetized and orotracheally intubated. The femoral artery was cannulated for pressure monitoring and blood sampling. The piglets were then placed in the prone position and mechanically ventilated for 2 d in the experimental intensive care unit (ICU) (19).

As shown in Figure 1, an Atomisor MegaHertz ultrasonic nebulizer (Diffusion Technique Française, Saint-Etienne, France), was positioned in the inspiratory limb, 40 cm proximal to the Y piece. After completion of nebulization, all parts of the ventilatory circuit were washed separately in a fixed volume of distilled water, and the amount of deposited amikacin was measured by an immunoenzymatic method. Before the in vivo study, the aerodynamic size distribution of particles was assessed in a separate bench study using a laser velocimeter.

View larger version (14K): [in this window] [in a new window]   Figure 1.   Diagram of the ventilator circuit. The nebulizer was positioned 40 cm before the Y piece. Nebulizer retention was defined as the amount of drug remaining in (A) after nebulization. Circuit deposition and endotracheal tube deposition were defined as the amount of drug trapped in (B) and (C). The exhaled dose was obtained by rinsing the expiratory filter (D).

Study Design

Eighteen piglets received 15 mg · kg¹ of amikacin as a single daily intravenous injection, and 18 received 1 g of amikacin powder diluted in 12 ml normal saline through ultrasonic aerosolization. During the 24 h after the first administration, blood and urine samples were collected. On Day 2, 24 h after the first administration, the second dose of amikacin was given and animals were killed at different time intervals for tissue sampling. Five pigs were killed 15 min after the second administration, five at 1 h, four at 3 h, and four at 6 h. Five additional piglets were killed 1 h after the first administration of amikacin to compare lung tissue concentrations after the first and the second administration.

Assessment of Amikacin Pharmacokinetics

In each animal, urinary and plasma amikacin concentrations were regularly measured after the first administration using an immunoenzymatic assay (Tdx, Abbott, IL). On Day 2, the second dose of amikacin was administered and piglets were killed at various intervals after drug administration. Lungs were exposed through a cervicothoracic incision, exsanguination was performed through direct cardiac puncture, and eight tissue specimens per animal (five subpleural and three juxtahilar) were collected from upper, middle, and lower lobes and stored at 20° C. Tissue samples were cryomixed in nitrogen, weighed, homogenized, and amikacin concentrations were measured by immunoenzymatic assay. Tissular hemoglobin, hematocrit, and plasma level of amikacin before sacrifice were measured, allowing the determination of the mass of amikacin resulting from the residual contaminating blood (20).

Pharmacokinetic Analysis

Standard kinetic parameters were determined. The analysis was performed with a software using a two-compartment (intravenous route) or a one-compartment (aerosol route) pharmacokinetic open model. In the aerosol group, the administered dose was defined as the amount of amikacin in the nebulizer minus the amount deposited in ventilatory circuits and expiratory filter.

Statistical Analysis

Data were analyzed using Statview software (SPSS, Inc., San Raphael, CA). Tissue and plasma amikacin concentrations were compared using a two-way analysis of variance for one within factor (time) and one grouping factor (intravenous versus aerosol administration). Differences in amikacin concentrations between lung segments and first versus second administration were analyzed using a nonparametric Wilcoxon test. A p value < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Measurement of Aerodynamic Size Distribution of Particles

For the three different locations (outlet of the nebulizer, closed to the Y piece and at the distal end of the endotracheal tube), mass median aerodynamic diameters were respectively 4.2 ± 0.2, 3.8 ± 0.3, and 3.4 ± 0.2 µm (mean ± SD g). Each value is the mean of six measurements (two measurements at each different time). The fraction of particles between 0.5 and 5 µm ranged from 45% to 50% at the distal tip of the endotracheal tube. No coalescence between particles was observed. The aerodynamic size distribution of particles was identical for the two tidal volumes tested, 340 and 600 ml.

Determination of Aerosol Extrapulmonary Deposition

Of the initial amount of amikacin placed in the nebulizer, 38% reached the bronchoalveolar compartment: 22 ± 6% was retained in the nebulizer's chamber and reservoir, 18 ± 8% was fixed in the inspiratory limb of the respiratory circuit, 4.5 ± 2% was deposited in the endotracheal tube, and 18 ± 7% deposited in the expiratory filter. Nebulization was completed in 20 min. The concentration of amikacin in the residual solution was higher than the initial concentration placed in the reservoir (120 mg · ml¹ versus 83 mg · ml¹).

Plasma Pharmacokinetics and Urinary Output

As shown in Figure 2, the mean plasma concentrations reached, respectively, 12 ± 6.6 µg · ml¹ 1.5 h after the end of the nebulization and 46 ± 6 µg · ml¹ 15 min after the end of the intravenous infusion. Plasma concentrations lower than 5 µg · ml¹ were measured after the sixth hour in both groups. A one-compartment open model with a first-order elimination rate and a two-compartment open model with first-order distribution and elimination rates adequately fitted the aerosol and intravenous plasma pharmacokinetics (the correlation coefficients between measured and predicted values were 93 ± 6% and 94 ± 6%). Similar elimination half-life values (4.6 ± 0.99 h for the aerosol and 4.1 ± 1 h for the intravenous route) were found for both groups. Figure 3 shows urinary elimination of aerosolized amikacin. The systemic bioavailability of aerosolized amikacin was 52 ± 19%. The ratio between the mean area under the curve of plasma concentrations (AUC_P) in the aerosol group and the mean AUC_P in the intravenous group was 0.58. 

View larger version (15K): [in this window] [in a new window]   Figure 2.   Mean amikacin plasma concentration time profiles after the end of administration by aerosol (closed circles, n = 10) or by intravenous infusion (open circles, n = 9). Data are expressed as mean ± SD.

View larger version (16K): [in this window] [in a new window]   Figure 3.   Fractionated urinary output of amikacin between 0 and 3, 3 and 6, 6 and 9, 9 and 12, 12 and 15, 15 and 18 h after nebulization of 1,000 mg of amikacin to 12 piglets. The mean cumulative urinary output of amikacin in the corresponding period of time is 207 ± 76 mg. Data are expressed as mean ± SD.

Lung Concentrations of Amikacin

Mean lung concentrations of amikacin at 15 min, 1 h, 3 h, and 6 h after the second administration were respectively 20, 12, 24, and 7-fold greater in the aerosol group than in the intravenous group (Figure 4). The area under the curve of amikacin lung concentrations over 6 h was 15-fold greater after aerosol administration than after intravenous infusion. These differences reached a high statistical significance (p < 0.0001). The distribution of nebulized amikacin showed no significant difference between lobes, juxtahilar and subpleural lung regions of each lobe (Figure 5), and between dependent and nondependent regions of lower lobes (175 ± 110 versus 165 ± 45 µg · g¹). No difference was found between right and left lungs. Lung concentrations of amikacin after intravenous administration were homogeneously distributed. Mean lung concentrations were not different between the group killed 1 h after the first aerosol and the group killed 1 h after the second aerosol (Figure 6). Lung concentrations of amikacin were corrected for blood contamination. Blood weight represented 4 ± 1% of lung samples, inducing an overestimation in lung concentrations of amikacin of 13 ± 19% after intravenous infusion and an underestimation of 11 ± 8% after aerosol administration.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Mean concentration of amikacin in lung tissue according to the time after the second aerosol administration (black bars) or the second intravenous infusion (gray bars). The second dose was given 24 h after the first dose. In each group (intravenous or aerosol), piglets were killed 15 min (n = 5), 1 h (n = 5), 3 h (n = 4), or 6 h (n = 4) after the administration. In each animal, five juxtapleural specimens were collected to obtain representative distal lung samples. Tissue concentrations of amikacin are corrected for blood contamination (see METHODS). Data are expressed as mean ± SD. AMK = amikacin. *p < 0.005.

View larger version (29K): [in this window] [in a new window]   Figure 5.   Regional distribution of lung tissue concentrations of nebulized amikacin 1 h after the second nebulization. Perihilar (P.H.) and juxtapleural tissue (J.P.) specimens were sampled in each lobe. Juxtapleural blocks sampled from lower lobe were excised from lung segment 10. In the lower lobe, two juxtapleural additional lung specimens were sampled from dependent (dep) and nondependent (nondep) lung regions, located respectively in segments 6 and 8. Data are expressed as mean ± SD. No significant statistical differences were found between the different tissue concentrations (Wilcoxon nonparametric test). AMK = amikacin.

View larger version (31K): [in this window] [in a new window]   Figure 6.   Lung tissue concentrations of amikacin 1 h after the first nebulization (black bars, n = 5) and 1 h after the second nebulization performed 24 h later (gray bars, n = 4). No significant difference was found. Data are expressed as mean ± SD.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study performed in piglets with healthy lungs shows that tissue concentrations of amikacin achieved in subpleural lung regions after aerosol administration were more than 10 times greater than those obtained after an intravenous administration. These striking differences in lung tissue concentrations were obtained although comparable amounts of amikacin reached the bronchoalveolar compartment after nebulization (38% of the initial amount deposited in the nebulizer) and the pulmonary capillary compartment after intravenous administration (100% of the intravenous dose). Tissue concentrations 10-fold greater than minimal inhibitory concentrations active against 90% of Pseudomonas aeruginosa and Acinetobacter baumannii strains were obtained 15 min after amikacin nebulization and maintained during 3 h. The pulmonary deposition of nebulized amikacin was homogeneously distributed, showing no parihilar to subpleural gradient of concentration. Tissue accumulation of amikacin could not be evidenced when comparing the tissue concentrations after two nebulizations at 24-h interval. Very low plasma concentrations of amikacin were measured after nebulization, and residual plasma levels  5 mmol · L¹ were observed after the sixth hour after the aerosol.

Aerosol Delivery

One of the factors explaining these high tissue concentrations is likely the optimization of the aerosol technique. An overall amount representing 40% of the initial nebulizer charge was delivered to the lungs. As rinsing of the circuit may not have collected the whole amount of amikacin that deposited, particularly considering the endotracheal tube which could be processed only at the end of the experiment, the extrapulmonary deposition may have been slightly underestimated. However, previous studies using radiolabeled aerosols reported a smaller lung deposition: in a series of mechanically ventilated patients, O'Riordan and coworkers reported a 15% (21) and a 22% (22) lung deposition using an optimized jet nebulizer. Three factors may have contributed to augment lung deposition in the present study: (1) the use of a large volume fill ultrasonic nebulizer (23) with a nebulizer retention representing less than 25% (24); (2) the use of an inspiratory time of 50% together with the interruption of the conventional humidifier (25); (3) the serial insertion of the nebulizer into the inspiratory limb, 40 cm before the Y piece (24), allowing an inspiratory "bolus" of the aerosol. During the expiratory phase, the aerosol accumulated in the inspiratory limb formed a bolus that was flushed into the airways during the subsequent inspiration.

Particles of 3.4 ± 0.2 µm internal diameter were obtained at the distal tip of the endotracheal tube, likely allowing a maximal distal airway deposition. Smaller particle sizes have been reported with different devices (25). However, in this study, true mass median aerodynamic diameter was possibly underestimated by cascade impactor measurement, which is less accurate than the laser particles sizer used in the present study. We found that particles were larger within the ventilator circuit than at the distal tip of the endotracheal tube, suggesting that there was a deposition of the largest particles in proximal parts of the ventilatory circuits. A tidal volume of 600 ml was also tested and had no influence on the aerodynamic size distribution. These data show that with an optimized aerosol technique, a substantial amount of amikacin may penetrate into the respiratory system using respiratory circuits and ventilatory settings commonly used in clinical practice.

Lung Tissue Concentrations after Amikacin Nebulization

In the present study, the immunoenzymatic method was used to measure lung tissue concentrations. Although this method was initially optimized for assays in plasma, it has also been widely used for assays in tissue (26) and in sputum (16). Tissue concentrations of amikacin were corrected for blood contamination. The measurement of tissue hemoglobin showed a small blood contamination because animals were exsanguinated before sampling. Nevertheless, the error resulting from blood contamination, depending on amikacin plasma level at the time of sacrifice, was substantial in the intravenous group, leading to an overestimation of 13 ± 19%, a value greater than usually observed (27).

At present, little information is available on lung tissue concentrations of nebulized antibiotics. A number of studies have investigated the concentrations of nebulized aminoglycosides in bronchial secretions, which is more indicative of the deposition in central airways than in distal bronchioles and lung parenchyma (11, 28). Other studies have evaluated the diffusion of tobramycin in the epithelial lining fluid through bronchoalveolar lavage (29, 30), which by diluting the alveolar sample, may markedly bias the true alveolar concentrations (30). Regarding lung tissue sampling, Leconte and coworkers reported high lung tissue concentrations 4 h or 12 h after the inhalation of tobramycin in patients undergoing thoracic surgery (31). A high variability of tissue concentrations was observed between patients, suggesting that the conditions of tobramycin administration and lung sampling were not standardized. For these reasons, we set up an animal model of prolonged mechanical ventilation mimicking the clinical conditions observed in critically ill patients and allowing a standardization of amikacin nebulization and pharmacokinetics assessment.

The site of penetration of aerosolized antibiotics remains uncertain. On peripheral lung samples, no distinction can be made between concentrations present in alveoli and bronchioles. It must be pointed out that more than 50% of nebulized amikacin was eliminated in urine. Considering the high degree of hydrophilia of amikacin resulting in a better absorption across the alveolo-capillary barrier than across the bronchial mucosa, these data suggest that a substantial amount of amikacin reached the alveolar space and subsequently diffused into the intravascular compartment. As aminoglycosides do not penetrate into cells, extracellular concentrations are likely higher than tissue concentrations actually measured in lung samples (32).

The lung regional distribution of amikacin (i.e., deposition in proximal versus peripheral airways) may be as important as the overall amount of lung deposition. No perihilar to juxtapleural gradient of tissue concentration was found in the present study. However, previous studies, using gamma techniques to characterize the pulmonary deposition of nebulized antimicrobial agents (16, 31, 33), found a preferential deposition in the proximal airways. In those areas, antibiotics may essentially deposit on the internal surface of the bronchial mucosa. Consequently, amikacin concentration measured in the perihilar lung specimens may have been underestimated, owing to the presence of tissues other than bronchial mucosa in the tissue homogenate.

Comparison between lung concentrations of amikacin measured 1 h after the first or the second aerosol did not show any cumulative effect, suggesting a low in situ residual concentration 24 h after the first administration. The very small elimination of amikacin in the last urine fraction (4.3 ± 3 mg) strengthens this hypothesis. Further studies involving repetitive daily administrations are required to assess a potential cumulative effect over a longer period of administration.

Lung Tissue Concentrations after the Intravenous Administration of Amikacin

The lung tissue concentrations after the administration of intravenous aminoglycosides have been assessed in several prior studies (6, 34). Most of these studies found tissue concentrations lower than 10 µg · ml¹ into sputum. Using a single daily administration of 15 mg · kg¹ in 10 critically ill patients with bronchopneumonia, Santre and coworkers (34) found a peak concentration of amikacin of 14 ± 9 µg · ml¹ in bronchial secretions, 3 h after the beginning of the infusion. In our study, a similar tissue peak concentration (13 ± 10 µg · g¹) occurred earlier, after 1 h, likely because of the time required for amikacin diffusion between lung tissue and bronchial secretions. After intravenous administration, lung tissue concentrations of amikacin reached levels in the range of minimal inhibitory concentrations of species such as P. aeruginosa, Citrobacter freundii, or A. baumannii, which are usually greater than 6 µg · L¹. As the antibacterial activity of amikacin appears concentration-dependent (3, 4), parenteral administration of amikacin may be much less efficient than nebulization for preventing or treating lung infection. Comparison between the first and the second daily intravenous dose was not performed, as the absence of cumulative effect after a single daily administration has been previously demonstrated (7, 34).

Plasma and Urine Pharmacokinetics

After inhalation, plasma concentrations of amikacin remained in the range of trough concentrations measured in the intravenous group and did not exceed 5 µg · L¹ from the sixth hour after administration. Because the lack of peak plasma concentrations after aerosolization very likely tended to restrict the tissue deposition in organs other than the lungs, a lower renal toxicity may be expected. Further studies measuring plasma concentrations after repetitive administrations are required because a potential cumulative effect may occur over a longer period of administration. Excluding the amount of amikacin that was trapped in the circuits or exhaled, 52 ± 19% of the amikacin available for the lung was eliminated in the urine during the first 18 h after aerosol. The remaining amount of drug should either be residual in the lungs, eliminated by mucociliary clearance, or fixed in other organs through blood perfusion. Because of its high hydrophilia, amikacin diffuses more easily across the alveolocapillary barrier than across the bronchial mucosa. As a consequence, after the inhalation of amikacin, its overall renal elimination could be a good index of its alveolar penetration. Compared with the nebulizer charge, the amount of drug collected in the urine (20.7 ± 7.6%) was much higher than found in previous studies (16, 31). The optimization of aerosol drug delivery resulting in a greater penetration into the distal lung parenchyma is likely the main determinant of this excellent bioavailability.

In conclusion, a substantial penetration of aerosolized amikacin into the distal lung of mechanically ventilated piglets was demonstrated using an optimized ultrasonic nebulizer. The lung concentrations of inhaled amikacin were more than 10-fold higher than the lung concentrations of intravenous amikacin, reaching levels far higher than the minimal inhibitory concentrations of the most resistant strains. Because this first study concerns healthy lungs, the main clinical implications may be the prevention of pulmonary infections as previously reported for intratracheal colistin (10) and the treatment of early stages of bronchopneumonia where the lung remains well aerated. The clinical relevance of the present data for treating confluent bronchopneumonia complicating mechanical ventilation cannot be extrapolated, as tissue concentrations of amikacin in infected and poorly aerated lung parenchyma may be lower. Further studies are required to investigate the lung deposition of nebulized amikacin in severely infected lung parenchyma.