- Split View
-
Views
-
Cite
Cite
Aukje C Bos, Johan W Mouton, Mireille van Westreenen, Eleni-Rosalina Andrinopoulou, Hettie M Janssens, Harm A W M Tiddens, Patient-specific modelling of regional tobramycin concentration levels in airways of patients with cystic fibrosis: can we dose once daily?, Journal of Antimicrobial Chemotherapy, Volume 72, Issue 12, December 2017, Pages 3435–3442, https://doi.org/10.1093/jac/dkx293
- Share Icon Share
Abstract
Inhaled tobramycin is important in the treatment of Pseudomonas aeruginosa (Pa) infections in cystic fibrosis (CF). However, despite its use it fails to attenuate the clinical progression of CF lung disease. The bactericidal efficacy of tobramycin is known to be concentration-dependent and hence changing the dosing regimen from a twice-daily (q12h) inhalation to a once-daily (q24h) inhaled double dose could improve treatment outcomes.
To predict local concentrations of nebulized tobramycin in the airways of patients with CF, delivered with the small airway-targeting Akita® system or standard PARI-LC® Plus system, with different inspiratory flow profiles.
Computational fluid dynamic (CFD) methods were applied to patient-specific airway models reconstructed from chest CT scans. The following q12h and q24h dosing regimens were evaluated: Akita® (150 and 300 mg) and PARI-LC® Plus (300 and 600 mg). Site-specific concentrations were calculated.
Twelve CT scans from patients aged 12–17 years (median = 15.7) were selected. Small airway concentrations were 762–2999 mg/L for the q12h dosing regimen and 1523–5997 mg/L for the q24h dosing regimen, well above the MIC for WT Pa strains. Importantly, the q24h regimen appeared to be more suitable than the q12h regimen against more resistant Pa strains and the inhibitory effects of sputum on tobramycin activity.
CFD modelling showed that high concentrations of inhaled tobramycin are indeed delivered to the airways, with the Akita® system being twice as efficient as the PARI-LC® system. Ultimately, the q24h dosing regimen appears more effective against subpopulations with high MICs (i.e. more resistant strains).
Introduction
Inhaled tobramycin is important in the treatment of Pseudomonas aeruginosa (Pa) infections in cystic fibrosis (CF) and is used to both eradicate early and suppress chronic Pa infections.1 As tobramycin exhibits concentration-dependent bactericidal activity, the highest possible concentrations above the MIC are required to be optimally efficacious against Pa.2 Furthermore, as sputum binding reduces its biological activity, inhaled concentrations in sputum must be at least 10-fold higher than the MIC for planktonic Pa and as high as 100–1000-fold greater for Pa growing in biofilm.3 Patients with CF are typically infected with multiple Pa morphotypes, each of which may vary in their susceptibility to antibiotics.4,5
Despite the use of current therapies to manage Pa infections, small airway disease in patients with CF continues to progress.6 This may be due to the insufficient deposition of inhaled antibiotics in the small airways, leading to local concentrations that are too low to be effective. Although high concentrations of tobramycin can be obtained in the central airways,7 concentrations in the small airways are yet unknown and difficult to measure in vivo.
Chronic treatment of Pa lung infection is currently defined as twice-daily tobramycin inhalation licensed with the PARI-LC® Plus nebulizer. Although this treatment regimen is used in all clinical trials, dosing once daily with a higher dose is likely to improve drug efficacy and safety. Specifically, higher peak levels could be obtained if the complete daily dose is delivered in a single inhalation. Recently, the pharmacokinetics of the once-daily inhaled double tobramycin dose were studied for the Akita® and PARI-LC® Plus nebulizers, where the Akita® is a smart nebulizer that allows for highly efficient targeting of the small airways. Similar pharmacokinetic profiles were found for both nebulizers and when compared with data on standard twice-daily inhalation, higher peak and lower trough levels were observed with the once-daily treatment regimen.8 As low trough levels are considered important to reduce aminoglycidal toxicity, safety may be improved by a longer clearance period and lower trough levels. Although promising, it is still unknown whether specific targeting of tobramycin to the small airways using a smart nebulizer results in sufficiently high antibiotic concentrations in the small airways.
By using airway models derived from CT scans of patients with CF who differ in disease severity, in combination with computational fluid dynamic (CFD) simulations, aerosol concentrations in the central and more distal airways can be computed. Furthermore, the relationship between airway morphology and airway concentrations can be assessed. This technique also yields similar information to single photon emission CT9 and has recently been used to study the association between structural lung disease in CF and the deposition of aztreonam lysine for inhalation (inhaled antibiotic for Pa treatment).10
Our study aimed to predict aerosol deposition patterns of inhaled tobramycin after once- and twice-daily dosing in CF lungs. We used patient-specific airway models and CFD simulations to determine the local concentrations per generation of the bronchial tree of inhaled tobramycin, delivered with the Akita® or PARI-LC® Plus nebulizer.
Methods
In a previous study, 40 patient-specific 3D models of the airways and lung lobes were reconstructed and used for CFD simulations.10 These airway models were reconstructed from spirometry controlled inspiratory and expiratory CT scans of patients with CF, acquired as part of routine clinical care. A total of 15 models were from children with CF aged ≥12 years. Of these 15 models we selected the six youngest and six oldest airway models for this study. All scans were anonymized and scored in random order by two experienced observers, whereby the validated CF-CT scoring system11 was used to quantify chest CT abnormalities. Component scores for bronchiectasis, airway wall thickening and mucus plugging were combined to compute the total airway disease (TAD) score and were expressed as a percentage of the maximum score (0%–100%).
Reconstruction of the airway models and simulations were performed by FLUIDDA NV (Kontich, Belgium) and have been previously reported.10 Briefly, the deposition of inhaled tobramycin was simulated for the Akita-Jet® compressor with the PARI-LC® Sprint nebulizer set to target peripheral airways, and for the Portaneb compressor with the PARI-LC® Plus nebulizer.
Ethics
Approval for this retrospective study was obtained from the Institutional Review Board of the Erasmus Medical Centre in Rotterdam, The Netherlands (MEC-2014-077). Written informed consent for the use of de-identified data was obtained from the parent/guardian and participants prior to inclusion in the study.
Reconstruction of 3D airway models
Automatic segmentation of the inspiratory scan was used to reconstruct a 3D model of the intra-thoracic region and could be performed down to the level of airways with a diameter of 1–2 mm. This was followed by a manual check of the airways; involving the addition of missing branches and the deletion of incorrect branches, as necessary.
For the CFD simulation, an upper airway model from an average adult was scaled down to match the average tracheal diameter at the location of the sternum for the paediatric population. This model was connected to the mouthpiece of the PARI-LC® Sprint nebulizer, which was further connected to the patient-specific lower airway model (Figure S1, available as Supplementary data at JAC Online).
Reconstruction of 3D lung lobes
To extract the patient-specific lung lobes from inspiratory and expiratory CT scans, a semi-automated tool that identifies the fissures separating the lobes was used (Mimics 15.0, Materialise N.V., Belgium, Food and Drug Administration, K073468; Conformité Européenne certificate, BE 05/1191.CE.01). For each of these lobes the volume change from expiration to inspiration was used to calculate the distribution of inhaled aerosol to that specific lobe.
Inlet of the airway model
Breathing profile and aerosol characteristics: Akita® nebulizer
The breathing profile for the simulations of tobramycin deposition with the Akita® was based on the predicted forced expiratory volume in 1 s (FEV1) value at time of the CT scan for each particular patient. For the 12 airway models, the inhalation time and volume per breath varied between 5 and 8 s and 1.0 and 1.6 L, respectively. The breathing profile had an inspiration–expiration ratio of 1:1.5 and was sinusoidal in shape. Tobramycin nebulization commenced immediately after inspiration began and continued up to the last second of the inhalation, after which a bolus of air was inhaled. The inhalation flow rate was fixed at 200 mL/s and a mass median aerosol diameter (MMAD) of 3.6 and geometric SD of 2.0 μm were used for inhalation of Bramitob by the Akita®.12
Loading doses of 150 and 300 mg of tobramycin were used in simulations for the Akita® system. Delivered doses are shown in Table 1 and were based on an in vitro study investigating the required delivered doses of different nebulizers to obtain an equivalent lung dose to the PARI-LC® Plus nebulizer.12,13 Owing to the efficiency of the Akita®, the delivered dose for this system was set to be much lower. However, the reduced loss of drug in the central airways meant that concentrations of drug in the small airways would be similar or perhaps even higher than the PARI-LC® Plus system. Additionally, both inspiration and expiration were modelled.
. | Akita® . | PARI-LC® Plus . |
---|---|---|
Loading dose twice daily | 150 | 300 |
Loading dose once daily | 300 | 600 |
Delivered dose twice daily | 26.95 | 96.2 |
Delivered dose once daily | 53.9 | 192.4 |
. | Akita® . | PARI-LC® Plus . |
---|---|---|
Loading dose twice daily | 150 | 300 |
Loading dose once daily | 300 | 600 |
Delivered dose twice daily | 26.95 | 96.2 |
Delivered dose once daily | 53.9 | 192.4 |
Loading doses and delivered doses used for the simulations. Doses are in milligrams. These doses are based on an in vitro study that aimed to calculate the required delivered dose with the Akita® nebulizer to obtain an equivalent lung dose to the PARI-LC® Plus nebulizer. Owing to the efficiency of Akita®, the delivered dose for this system was set to be much lower to obtain an equivalent lung dose to PARI-LC® Plus. Both nebulizers in this study were filled with 300 mg of tobramycin.12,13 Thus, for Akita® these doses belong to the once-daily regimen in our study and for PARI-LC® Plus these doses belong to the twice-daily regimen.
. | Akita® . | PARI-LC® Plus . |
---|---|---|
Loading dose twice daily | 150 | 300 |
Loading dose once daily | 300 | 600 |
Delivered dose twice daily | 26.95 | 96.2 |
Delivered dose once daily | 53.9 | 192.4 |
. | Akita® . | PARI-LC® Plus . |
---|---|---|
Loading dose twice daily | 150 | 300 |
Loading dose once daily | 300 | 600 |
Delivered dose twice daily | 26.95 | 96.2 |
Delivered dose once daily | 53.9 | 192.4 |
Loading doses and delivered doses used for the simulations. Doses are in milligrams. These doses are based on an in vitro study that aimed to calculate the required delivered dose with the Akita® nebulizer to obtain an equivalent lung dose to the PARI-LC® Plus nebulizer. Owing to the efficiency of Akita®, the delivered dose for this system was set to be much lower to obtain an equivalent lung dose to PARI-LC® Plus. Both nebulizers in this study were filled with 300 mg of tobramycin.12,13 Thus, for Akita® these doses belong to the once-daily regimen in our study and for PARI-LC® Plus these doses belong to the twice-daily regimen.
Breathing profile and aerosol characteristics: PARI-LC® Plus nebulizer
The sinusoidal breathing profiles for the simulations of tobramycin deposition with the PARI-LC® Plus had an inspiration–expiration ratio of 1:1.5, where tobramycin nebulization was continuous during both inspiration and expiration. Furthermore, these profiles were generated using the age and height of each patient at time of the CT scan, and the reference formula developed by Zapletal et al.14 Specifically, the mean, upper and lower limits of this formula were used to generate a mean, high and low tidal volume, respectively. Several trials have studied the diameter distribution of tobramycin nebulized with the PARI-LC® Plus system in combination with different compressors. From these studies the smallest and largest reported MMAD were used for the CFD simulations: smallest MMAD (3.4 μm; TOBI® nebulized with PARI-LC® Plus combined with Turboboy SX compressor)15 and largest MMAD (4.93 μm; TOBI® nebulized with PARI-LC® Plus combined with DevilBiss PulmoAide compressor).16 Additionally, a geometric SD of 2.3 was used in these simulations.17
Loading doses of 300 and 600 mg of tobramycin were used in simulations for the PARI-LC® Plus system. Delivered doses are shown in Table 1 and both inspiration and expiration were modelled. Unlike the Akita®, circulating particles that were neither deposited nor exhaled were present following the first expiration with the PARI-LC® Plus. Therefore, the respiratory cycle was simulated twice, from which the results of the second cycle were used for computation of the concentrations.
Computation of tobramycin concentrations
To compute regional tobramycin deposition, the airway surface area of the respiratory tract was subdivided into two regions. For airways with a diameter >1–2 mm (i.e. large airways), tobramycin concentrations were computed using the combined mouthpiece–upper/lower airway model derived from chest CT scans.
Airways with a diameter <1–2 mm (i.e. small airways) were generally not visible on the CT images and hence were added to the model using Phalen’s description of the airway tree in infants, children and adolescents.18 Phalen’s data were obtained from subjects without lung disease and describe airway dimensions up to the 16th generation. Once the aerosol entered a lobe in the Phalen section of the model, it was assumed that it would be distributed homogeneously. Regional peak tobramycin concentrations in the epithelial lining fluid (ELF) were computed as follows: the fraction of deposited inhaled aerosol in an airway multiplied by the delivered dose was divided by the surface area of that airway multiplied by the ELF thickness. A range of ELF thickness was used and based on studies in CF (3, 5 and 7 μm).19 Tobramycin concentrations were calculated for the following variables: the Akita® (one MMAD and one breathing profile) or PARI-LC® Plus (two MMADs and three tidal volumes) system, ELF thickness (3, 5 and 7 μm) and the dosing regimen (once- and twice-daily doses). Results are described for the median ELF thickness of 5 μm unless otherwise indicated.
Statistical analysis
Patient characteristics and tobramycin concentrations were summarized using descriptive statistics and all data are presented as the median (IQR). Concentrations described in the text refer to the median. IQRs can be found in the tables.
The mixed-effects model was used to assess differences in concentrations between the simulated situations and the significance level was set to 0.05. Specifically, the concentrations calculated for the median ELF were used as outcomes and a separate analysis was performed for the once- and twice-daily dosing regimens. The following variables were included in the model as predictors: TAD score, percentage predicted FEV1 (FEV1%pred) and percentage predicted forced expiratory flow at 75% of forced vital capacity (FEF75%pred). The advantage of using mixed-effects models is that they account for repeated measurements on the same patients.
Descriptive statistics and intra-class correlation coefficients were calculated using SPSS/PC Statistics 21.0 (SPSS Inc., Chicago, IL, USA). Mixed-effects modelling was performed with the statistical software package R (free download from www.rproject.org) version 3.2.2.
Results
Study population
Twelve spirometry controlled inspiratory and expiratory chest CT scans (n = 8 female) were selected for this study. However, two airway models from the same patient were included from two different timepoints, 2 years apart. Hence in total the airway models were derived from 11 patients. Baseline characteristics are shown in Table 2 and disease severity, as indicated by the TAD score and FEV1%pred, was highly variable.
Male | 33.3% |
Age (years), median (IQR) | 15.7 (13.9–17.1) |
TAD score (percentage of the maximum score), median (IQR) | 7.3 (4.1–11.1) |
FEV1%pred, median (IQR) | 92.6 (86.4–115.9) |
FEF75%pred, median (IQR) | 54.7 (35.2–71.5) |
Male | 33.3% |
Age (years), median (IQR) | 15.7 (13.9–17.1) |
TAD score (percentage of the maximum score), median (IQR) | 7.3 (4.1–11.1) |
FEV1%pred, median (IQR) | 92.6 (86.4–115.9) |
FEF75%pred, median (IQR) | 54.7 (35.2–71.5) |
Male | 33.3% |
Age (years), median (IQR) | 15.7 (13.9–17.1) |
TAD score (percentage of the maximum score), median (IQR) | 7.3 (4.1–11.1) |
FEV1%pred, median (IQR) | 92.6 (86.4–115.9) |
FEF75%pred, median (IQR) | 54.7 (35.2–71.5) |
Male | 33.3% |
Age (years), median (IQR) | 15.7 (13.9–17.1) |
TAD score (percentage of the maximum score), median (IQR) | 7.3 (4.1–11.1) |
FEV1%pred, median (IQR) | 92.6 (86.4–115.9) |
FEF75%pred, median (IQR) | 54.7 (35.2–71.5) |
Deposition analyses
High concentrations of inhaled tobramycin were delivered to all regions of the lungs by both nebulizers. For the Akita®, the median tobramycin concentration in the large airways was 77 255 mg/L when dosed twice daily and 154 511 mg/L when dosed once daily. For the PARI-LC® Plus, median tobramycin concentrations in the large airways, calculated for high to low tidal volume and both MMADs, were 84 316–94 957 mg/L when dosed twice daily and 168 633–189 916 mg/L when dosed once daily.
Figure 1 shows tobramycin concentrations in the small airways calculated for the different lining fluid thicknesses. For the median ELF (Figure 1b), simulations with the Akita® resulted in median tobramycin concentrations of 1770 mg/L when dosed twice daily and 3541 mg/L when dosed once daily. For the PARI-LC® Plus, median tobramycin concentrations were 1066–1800 mg/L when dosed twice daily and 2133–3598 mg/L when dosed once daily, again calculated for high to low tidal volume and both MMADs. Figure 1 also shows how tidal volume and MMAD influence tobramycin concentrations in the small airways for the PARI-LC® Plus. Specifically, lower tobramycin concentrations in the small airways were observed with higher tidal volumes and the largest MMAD.
Significantly lower tobramycin concentrations were observed with the PARI-LC® Plus nebulizer than with the Akita® nebulizer, with the exception of inhalation at a low breathing tidal volume and the smallest MMAD (3.4 μm) (Table 3). For example, if two patients with the same TAD score, FEV1%pred and FEF75%pred were compared, the patient using the PARI-LC® Plus with a low tidal volume and an aerosol diameter of 4.93 μm will have a mean reduction in the peripheral tobramycin concentration of 652 and 326 mg/L (once and twice daily, respectively) compared with the Akita®. Much smaller and reverse associations were seen for FEF75%pred and the TAD score.
Characteristic/ nebulizer . | MMAD (μm) . | Concentration (mg/L) . | Estimate . | Standard error of estimate . | P . | |||
---|---|---|---|---|---|---|---|---|
once daily . | twice daily . | once daily . | twice daily . | once daily . | twice daily . | |||
FEV1%pred | – | – | – | 3.1 | 1.6 | 1.7 | 0.8 | 0.062 |
FEF75%pred | – | – | – | 4.3 | 2.1 | 0.9 | 0.4 | <0.001 |
TAD score | – | – | – | 12.8 | 6.4 | 3.8 | 1.9 | <0.001 |
Akita® | 3.6 | 3540.94 (3327.21–3784.47) | 1770.47 (1663.61–1892.23) | |||||
PARI-LC® Plus (low tidal volume) | 3.4 | 3598.48 (3367.05–3651.84) | 1799.79 (1683.52–1825.92) | 28.3 | 14.1 | 74.5 | 37.3 | 0.704 |
4.93 | 2928.16 (2718.61–3008.46) | 1464.08 (1359.31–1504.23) | −651.9 | −326.0 | 72.0 | 36.0 | <0.001 | |
PARI-LC® Plus (mean tidal volume) | 3.4 | 3284.25 (3093.91–3351.16) | 1642.13 (1546.95–1675.58) | −288.5 | −144.3 | 70.0 | 35.0 | <0.001 |
4.93 | 2562.62 (2372.88–2615.50) | 1281.31 (1186.44–1307.75) | −1013.3 | −506.6 | 66.9 | 33.4 | <0.001 | |
PARI-LC® Plus (high tidal volume) | 3.4 | 2871.36 (2725.62–3003.25) | 1435.68 (1362.81–1501.62) | −659.2 | −329.6 | 66.3 | 33.2 | <0.001 |
4.93 | 2132.66 (2008.99–2288.59) | 1066.33 (1004.49–1144.29) | −1393.7 | −696.8 | 64.2 | 32.1 | <0.001 |
Characteristic/ nebulizer . | MMAD (μm) . | Concentration (mg/L) . | Estimate . | Standard error of estimate . | P . | |||
---|---|---|---|---|---|---|---|---|
once daily . | twice daily . | once daily . | twice daily . | once daily . | twice daily . | |||
FEV1%pred | – | – | – | 3.1 | 1.6 | 1.7 | 0.8 | 0.062 |
FEF75%pred | – | – | – | 4.3 | 2.1 | 0.9 | 0.4 | <0.001 |
TAD score | – | – | – | 12.8 | 6.4 | 3.8 | 1.9 | <0.001 |
Akita® | 3.6 | 3540.94 (3327.21–3784.47) | 1770.47 (1663.61–1892.23) | |||||
PARI-LC® Plus (low tidal volume) | 3.4 | 3598.48 (3367.05–3651.84) | 1799.79 (1683.52–1825.92) | 28.3 | 14.1 | 74.5 | 37.3 | 0.704 |
4.93 | 2928.16 (2718.61–3008.46) | 1464.08 (1359.31–1504.23) | −651.9 | −326.0 | 72.0 | 36.0 | <0.001 | |
PARI-LC® Plus (mean tidal volume) | 3.4 | 3284.25 (3093.91–3351.16) | 1642.13 (1546.95–1675.58) | −288.5 | −144.3 | 70.0 | 35.0 | <0.001 |
4.93 | 2562.62 (2372.88–2615.50) | 1281.31 (1186.44–1307.75) | −1013.3 | −506.6 | 66.9 | 33.4 | <0.001 | |
PARI-LC® Plus (high tidal volume) | 3.4 | 2871.36 (2725.62–3003.25) | 1435.68 (1362.81–1501.62) | −659.2 | −329.6 | 66.3 | 33.2 | <0.001 |
4.93 | 2132.66 (2008.99–2288.59) | 1066.33 (1004.49–1144.29) | −1393.7 | −696.8 | 64.2 | 32.1 | <0.001 |
Differences in concentrations between the simulated situations. Associations between small airway tobramycin concentrations [median (IQR)] and disease characteristics (TAD score, FEV1%pred and FEF75%pred) or nebulizer with breathing pattern for the median ELF and different aerosol diameters. For concentrations calculated for the thin and thick ELF we refer to Figure 1(a and c). Comparison is made with the Akita® nebulizer. For example: if two patients with the same disease severity (same TAD score, FEV1%pred and, FEF75%pred) were compared, the patient using PARI-LC® Plus with a low tidal volume and an aerosol diameter of 4.93 μm will have a mean reduction in the peripheral tobramycin concentration of 652 and 326 mg/L (once and twice daily, respectively) compared with Akita®. P values in bold represent significant differences.
Characteristic/ nebulizer . | MMAD (μm) . | Concentration (mg/L) . | Estimate . | Standard error of estimate . | P . | |||
---|---|---|---|---|---|---|---|---|
once daily . | twice daily . | once daily . | twice daily . | once daily . | twice daily . | |||
FEV1%pred | – | – | – | 3.1 | 1.6 | 1.7 | 0.8 | 0.062 |
FEF75%pred | – | – | – | 4.3 | 2.1 | 0.9 | 0.4 | <0.001 |
TAD score | – | – | – | 12.8 | 6.4 | 3.8 | 1.9 | <0.001 |
Akita® | 3.6 | 3540.94 (3327.21–3784.47) | 1770.47 (1663.61–1892.23) | |||||
PARI-LC® Plus (low tidal volume) | 3.4 | 3598.48 (3367.05–3651.84) | 1799.79 (1683.52–1825.92) | 28.3 | 14.1 | 74.5 | 37.3 | 0.704 |
4.93 | 2928.16 (2718.61–3008.46) | 1464.08 (1359.31–1504.23) | −651.9 | −326.0 | 72.0 | 36.0 | <0.001 | |
PARI-LC® Plus (mean tidal volume) | 3.4 | 3284.25 (3093.91–3351.16) | 1642.13 (1546.95–1675.58) | −288.5 | −144.3 | 70.0 | 35.0 | <0.001 |
4.93 | 2562.62 (2372.88–2615.50) | 1281.31 (1186.44–1307.75) | −1013.3 | −506.6 | 66.9 | 33.4 | <0.001 | |
PARI-LC® Plus (high tidal volume) | 3.4 | 2871.36 (2725.62–3003.25) | 1435.68 (1362.81–1501.62) | −659.2 | −329.6 | 66.3 | 33.2 | <0.001 |
4.93 | 2132.66 (2008.99–2288.59) | 1066.33 (1004.49–1144.29) | −1393.7 | −696.8 | 64.2 | 32.1 | <0.001 |
Characteristic/ nebulizer . | MMAD (μm) . | Concentration (mg/L) . | Estimate . | Standard error of estimate . | P . | |||
---|---|---|---|---|---|---|---|---|
once daily . | twice daily . | once daily . | twice daily . | once daily . | twice daily . | |||
FEV1%pred | – | – | – | 3.1 | 1.6 | 1.7 | 0.8 | 0.062 |
FEF75%pred | – | – | – | 4.3 | 2.1 | 0.9 | 0.4 | <0.001 |
TAD score | – | – | – | 12.8 | 6.4 | 3.8 | 1.9 | <0.001 |
Akita® | 3.6 | 3540.94 (3327.21–3784.47) | 1770.47 (1663.61–1892.23) | |||||
PARI-LC® Plus (low tidal volume) | 3.4 | 3598.48 (3367.05–3651.84) | 1799.79 (1683.52–1825.92) | 28.3 | 14.1 | 74.5 | 37.3 | 0.704 |
4.93 | 2928.16 (2718.61–3008.46) | 1464.08 (1359.31–1504.23) | −651.9 | −326.0 | 72.0 | 36.0 | <0.001 | |
PARI-LC® Plus (mean tidal volume) | 3.4 | 3284.25 (3093.91–3351.16) | 1642.13 (1546.95–1675.58) | −288.5 | −144.3 | 70.0 | 35.0 | <0.001 |
4.93 | 2562.62 (2372.88–2615.50) | 1281.31 (1186.44–1307.75) | −1013.3 | −506.6 | 66.9 | 33.4 | <0.001 | |
PARI-LC® Plus (high tidal volume) | 3.4 | 2871.36 (2725.62–3003.25) | 1435.68 (1362.81–1501.62) | −659.2 | −329.6 | 66.3 | 33.2 | <0.001 |
4.93 | 2132.66 (2008.99–2288.59) | 1066.33 (1004.49–1144.29) | −1393.7 | −696.8 | 64.2 | 32.1 | <0.001 |
Differences in concentrations between the simulated situations. Associations between small airway tobramycin concentrations [median (IQR)] and disease characteristics (TAD score, FEV1%pred and FEF75%pred) or nebulizer with breathing pattern for the median ELF and different aerosol diameters. For concentrations calculated for the thin and thick ELF we refer to Figure 1(a and c). Comparison is made with the Akita® nebulizer. For example: if two patients with the same disease severity (same TAD score, FEV1%pred and, FEF75%pred) were compared, the patient using PARI-LC® Plus with a low tidal volume and an aerosol diameter of 4.93 μm will have a mean reduction in the peripheral tobramycin concentration of 652 and 326 mg/L (once and twice daily, respectively) compared with Akita®. P values in bold represent significant differences.
Discussion
In this study, CFD was used to compute local inhaled tobramycin concentrations throughout the bronchial tree in CF after once- and twice-daily dosing. High concentrations of inhaled tobramycin were delivered to all lung regions, with the Akita® nebulizer being twice as efficient as the PARI-LC® Plus. The average small airway concentrations calculated for the twice-daily dose (median lining fluid, both nebulizers) correlated well with previously published data, as tobramycin concentrations up to ∼1750 mg/L were measured in sputum.2,20 The once-daily dose resulted in higher tobramycin concentrations in the small airways compared with the twice-daily dose (1523–5997 versus 762–2999 mg/L, respectively; both nebulizers combined). This result is promising due to the concentration-dependent bactericidal efficacy of tobramycin; where, the higher the concentration, the greater the reduction in bacterial density.2
However, whether the computed concentrations are sufficient for effective killing throughout the lung remains questionable, as the concentration at which effective killing is obtained is as yet unknown. Owing to the substantial heterogeneity of phenotypes and genotypes of Pa within a single patient, a range of MIC values exist throughout the lung.21 Thus, the in vitro MIC value does not reflect the wide ranges of MICs that can be present within the lungs of a single patient. Additionally, the MIC reflects the activity of a drug under specific optimal circumstances, and may be far less for microorganisms in a semi-dormant state. Furthermore, in vitro measurements do not take into account the hostile lung environment for antibiotics, where the activity of inhaled tobramycin in the lungs is reduced due to binding to mucin and DNA fragments within the mucus22,23 and due to biofilm formation by Pa.24 To overcome the effect of sputum binding on tobramycin, it is generally assumed that local concentrations need to be 10–25-fold above the MIC. WT organisms refer to the phenotype of the typical form of a species, as they occur in nature. These organisms may have intrinsic resistance mechanisms, but not acquired mechanisms. Pseudomonas WTs have an MIC of ≤2 mg/L.25,26 Hence for WT Pa (MIC2 mg/L), effective killing can be easily achieved as concentrations 381–1500 × MIC2 mg/L were observed in the small airways for the twice-daily dose and 762–2999× MIC2 mg/L for the once-daily dose. However, for Pa strains with in vitro MICs >2 mg/L (i.e. acquired resistance mechanisms), the concentrations required for effective killing of these strains remain unknown. Assuming the highest MIC value measured in vitro, 512 mg/L (MIC512 mg/L), only the once-daily dose calculated for a thin lining fluid resulted in concentrations 10-fold greater than the MIC (10× MIC512 mg/L). For these simulations, all patients received small airway concentrations >10× MIC512 mg/L for both the Akita® and PARI-LC® Plus systems when nebulized with a low tidal volume and small aerosol diameter. For other simulations, only a proportion of the patients received concentrations above the threshold in all airways for the PARI-LC® Plus (Figure 2).
The problems associated with high MICs and additional sputum binding may be partly overcome by increasing the tobramycin concentration, which serves as the rationale for administering a double dose of tobramycin in a single inhalation. Higher tobramycin concentrations will ultimately result in increased concentrations of free drug that are able to kill Pa.27 Likewise, higher tobramycin concentrations will allow antibiotic particles to penetrate deeper into bacterial microcolonies, as the diffusion of tobramycin is concentration-dependent. The optimal drug concentration will differ between patients, as there is marked variability between patients in the inhibitory activity of mucus against the killing efficacy of aminoglycosides such as tobramycin. Moreover, mucus layer thickness varies between patients and throughout the bronchial tree. To overcome the antagonistic activity of mucus in all patients, including those with the worst case sputum composition, peak tobramycin concentrations had to be 100 times the MIC to ensure killing of planktonic Pa,28 or even 100–1000-fold greater for Pa growing in biofilm.3 This means that patients with greater antagonistic mucus activity and highly resistant Pa strains would possibly need even higher concentrations than the once-daily dose simulated in our study.
Another reason for once-daily dosing of an increased dose of inhaled tobramycin is that aminoglycosides induce a post-exposure effect and therefore need to be dosed less frequently than β-lactam antibiotics, for example. Tobramycin exposure induces sublethal damage in Pa bacteria, which needs to be repaired before regrowth can commence to allow a new dose of aminoglycosides to be effective. The time it takes to repair this damage in part correlates with the post-exposure effect and continues when the antibiotic concentration falls below the MIC. A post-exposure effect of ∼2 h has been described for aminoglycosides against Pa in vitro using an enzymatic inactivation method. During a simulation in mice, the in vivo post-antibiotic effect was even longer (∼5 h) than the effect in vitro,29 as longer half-lives increase the duration of this effect.30 The half-life of tobramycin measured in sputum was ∼2 h post-nebulization of 80 mg of tobramycin.31
Although our study supports the use of the once-daily treatment regimen of nebulized tobramycin, drug safety issues need to be considered when increasing nebulized doses of aminoglycosides. A pharmacokinetic study in adult patients with CF showed that inhalation of a double tobramycin dose with either the Akita® or PARI-LC® Plus was well tolerated and resulted in higher peak levels (i.e. likely improved antibiotic effect), though trough levels remained well below the toxic limit.8 Additionally, studies of intravenous tobramycin showed that a longer clearance period of systemically absorbed tobramycin is associated with a better safety profile,32,33 suggesting that a once-daily, double dose of inhaled tobramycin is safe.8 Future clinical studies are needed to examine if once-daily inhalation of a double tobramycin dose is more effective in patients with CF infected with Pa, and whether chronic use of this treatment regimen is safe.
A previous CFD study investigating concentrations of aztreonam lysine for inhalation in relation to structural lung disease in patients with CF showed that the more diseased lobes received lower levels of the inhaled antibiotic,10 and this is important for inhaled tobramycin due to its low rate of systemic absorption (9%–17.5%).34 This means that hypoventilated lung areas are undertreated as systemic absorption contributes little to the treatment effect of these areas. For antibiotics with relatively good absorption from the lung (e.g. inhaled levofloxacin),21 a clinical response in these hypoventilated areas is likely to occur following systemic absorption via the bronchial circulation.
The limitations of CFD modelling have been previously described.10 First, to simulate and predict tobramycin concentrations, assumptions had to be made for ELF thickness and effective tobramycin concentrations. Secondly, the model does not account for sputum binding or mucociliary clearance. Finally, the following parameters were not patient-specific: the upper airway, airways with a diameter <1–2 mm, and breathing profiles. An exception to this limitation would be the breathing profiles for the Akita®, which were based on the FEV1 value of that specific patient similar to what would be used in clinics. For the PARI-LC® Plus, a range of breathing profiles was simulated based on the age and height of the specific patient. Future modelling studies could implement patient-specific breathing profiles and upper airways to make an even more reliable estimation of the required tobramycin concentrations for specific patients.
Conclusions
Our model predicts that high concentrations of inhaled tobramycin are delivered to the small airways of the lungs, with the Akita® being twice as efficient as the PARI-LC® Plus. For effective killing of more resistant Pa strains, inhalation of a double tobramycin dose would be required. As inhalation of a double dose is not associated with acute toxicity, we thus recommend a once-daily, double dose of nebulized tobramycin in patients who do not improve with standard care. The Akita® would be more efficient as only half the dose is required, which reduces treatment time. However, the PARI-LC® Plus could also be used when the patient is able to execute a slow and deep inhalation throughout nebulization.
Acknowledgements
Data from the manuscript have been presented as an e-poster at the Thirty-ninth European Cystic Fibrosis Society Conference (ECFS), Basel, Switzerland, 2016 (ePS06.4).
Funding
This study was supported by an unconditional grant from Chiesi Farmaceutici S.p.A. This was an investigator-initiated study.
Transparency declarations
J. W. M. has received research funding from Adenium, AstraZeneca, Basilea, Cubist, Polyphor, Roche, Eumedica, Basilea, VenatorX, AiCuris, Gilead and Wockhardt. H. M. J. reports incidental consultancies from Vertex and Gilead, outside the submitted work. All financial aspects of the above mentioned activities are handled by the BV Kindergeneeskunde of the Erasmus MC-Sophia Children s Hospital. H. A. W. M. T. reports other from Roche, other from Pharmaxis, other from Novartis, grants from CFF, grants and other from Vertex and grants and other from Gilead, outside the submitted work. In addition, H. A. W. M. T. has a patent licenced with Activaero, has a patent PRAGMA-CF scoring system issued and is heading the Erasmus MC-Sophia Children's Hospital core laboratory LungAnalysis. All financial aspects of the above-mentioned activities are handled by the BV Kindergeneeskunde of the Erasmus MC-Sophia Children's Hospital. All other authors: none to declare.
Supplementary data
Figure S1 is available as Supplementary data at JAC Online.