Dataset Viewer
institute
stringclasses 10
values | file
stringlengths 7
154
| pseudo_pagenum
int64 0
1.15k
| text
stringlengths 102
7.59M
| Institute
stringclasses 10
values |
|---|---|---|---|---|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 0
|
General introduction InTroduCTIon, AIms & ouTlInE Cystic fibrosis Cystic fibrosis (CF) is a severe hereditary and life-threatening disease in the Caucasian population, affecting 70,000 patients worldwide. 1 In the 1950s, a child with CF would rarely live long enough to attend elementary school. Luckily, life expectancy has dramati- cally improved due to the development of new treatments and treatment approaches. The current median predicted survival is close to 40.The main cause of mortality and also morbidity is progressive lung disease. 3 Patients with CF have an increased susceptibility to airway infections due to the defects in the CF transmembrane conductance regulator (CFTR). Due to a dysfunction in the CFTR channel, ion transfer over the membrane of epithelial cells is abnormal what leads to dehydration of the airway surface liquid in the lungs. This results in thick, viscous mucus that impairs mucociliary clearance and obstructs the airways. Because of these alterations, it is dif - ficult for CF patients to eliminate inhaled bacteria from the lungs. A persistent infection arises, which damages the airways and eventually leads to respiratory failure.The most common mutation causing CFTR dysfunction is the dF508 mutation, but over 2000 mutations have been identified.1 The disease is complex and the rate of progression of CF lung disease varies widely from person to person, even among patients carrying identical CFTR mutations.4 Despite the significant improvements in the treatment of the disease, there is still no cure and lung infections remain a serious problem for patients living with CF. Small airways disease An important component in the pathophysiology of CF lung disease is small airways disease,5-7 which starts early in life. Most infants with CF diagnosed through newborn screening have evidence of small airways disease already in the first year of life.8 Differ- ences in infant lung function are seen between children with CF and healthy control sub- jects and areas of bronchiectasis, mucus plugging and air trapping are observed on chest computed tomography (CT) scans. 4 In addition, in patients with end stage lung disease 10-75% of the lung volume is dysfunctional due to small airways disease.9 Despite this large body of evidence that small airways disease is an important component of CF lung disease, current aerosol therapy is probably inefficient in targeting these small airways. Treatment of CF lung disease The mainstay of the management of lung disease in patients with CF is aerosol therapy. Inhaled drugs are used to thin the mucus or fight lung infections. Multiple microorganisms play a role in the pulmonary infections in CF patients, but the most important contributor to progression of CF lung disease is Pseudomonas aeruginosa (Pa). 10 Inhaled antibiot -
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 1
|
Chapter ics play a key role in the eradication and chronic suppressive therapy of Pa infections. Currently, 3 inhaled antibiotics are registered for use in patients with CF: Tobramycin, Aztreonam Lysine and Colistin. Inhaled antibiotics First attempts of inhaling antibiotics, 50 years ago, were done by off label use of par - enteral preparations in conventional nebulizers. However, due to the osmolality and preservatives in the preparations their use was not tolerated very well.Tobramycin In 1997 tobramycin inhalation solution (TIS) came to market, especially designed for the lung as a preservative free formulation that had an osmolality and pH closer to that of the airway surface liquid.1,11 In large clinical studies it was shown that maintenance with TIS significantly increased lung function up to 12%, reduced bacterial density in sputum and reduced pulmonary exacerbations.1 Since its FDA approval it is the first choice of therapy for lung infections caused by Pa. Tobramycin is an aminoglycoside antibiotic and works by inhibiting protein synthesis of the bacteria. For the treatment of Pa lung infections twice daily nebulization of 300 mg tobramycin during 28 days is recommended in a month on month off regimen. However, various dosage regimens were investigated, with dosages ranging from 80 mg twice daily to 600 mg thrice daily. 12,13 For a first infection with Pa, a one month treatment with TIS is sufficient to eradicate Pa in two thirds of the patients. Unfortunately, this still means that in around one third of patients eradication therapy of Pa fails and the infection becomes chronic.14 One of the possibilities for failure of TIS to eradicate Pa is that its bactericidal effect is concentration dependent. Thus insufficient peak concentrations throughout the lung will result in incomplete killing of Pa. Aztreonam lysine for inhalation In 2010, aztreonam lysine for inhalation (AZLI) was approved by the FDA for treatment of Pa infections. AZLI is a monobactam antibiotic that has been proven to improve lung function and reduce respiratory symptoms, bacterial density and exacerbations. It inhibits cell wall biosynthesis of gram-negative bacteria. Instead of aminoglycosides, AZLI demonstrates time-dependent killing. Thus, the degree of bacterial killing does not primarily depend on the peak concentration of the antibiotic, but is more dependent on the total time that AZLI concentrations remain above the susceptibility breakpoint of bacteria.15 The recommended dosing regimen for AZLI is three times daily 75 mg, like TIS in a month on month off regimen.1
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 2
|
General introduction Colistin Nebulized colistin is frequently used for treatment of Pa lung infections in patients with CF despite very few controlled trials evaluated its efficacy. For eradication of early Pa infections, only the combination of nebulized colistin with oral ciprofloxacin has been investigated. Compared to no treatment, eradication of Pa was seen more often in the colistin group (odds ratio 0.12) (26 children). Two trials compared nebulized colistin with oral ciprofloxacin to either TIS alone (58 children) or TIS with oral ciprofloxacin (223 patients >1 year) and showed comparable efficacy to TIS. 16 For chronic Pa infections, 2 trials compared nebulized colistin to placebo (54 patients in total) and 1 trial compared colistin to TIS (115 patients). No advantage was seen in patients treated with colistin compared to placebo. In the comparison to TIS, patients treated with colistin did not show any improvement in lung function while an increase of 6.7% in lung function was seen in patients receiving TIS.3 Colistin is a polymyxin antibiotic and works by binding to and disrupting the outer cell membrane of bacteria. One million units colistin (=80 mg) are administered twice daily. Dry powder inhalers More recently, dry powder formulations of antibiotics have been developed as a time ef- ficient alternative for nebulized antibiotics. Administration is much quicker, for example for tobramycin the podhaler was developed which allows inhaling the required dose in 5 min instead of approximately 20 minutes when inhaled by nebulizer. Clearly, this can greatly improve adherence to therapy. Colobreathe is a dry powder inhaler for colistin. Both dry powders showed noninferiority relative to nebulized tobramycin. 17,18 A disad- vantage of dry powder inhalers is that the aerosol characteristics of the inhaled drug are dependent on the inhalation maneuver by the patient. In this thesis we focus on inhaled antibiotics administered by nebulizers. Other inhaled drugs in CF Other inhaled drugs used by patients with CF are mucolytics/mucous mobilizers, anti- inflammatory drugs and bronchodilators. Apart from dornase alfa, these inhaled drugs are outside the scope of this thesis. Dornase alfa Dornase alfa (Pulmozyme) is a highly purified solution of recombinant human deoxyri - bonuclease (rhDNase). It reduces viscoelasticity of CF sputum by cleaving extracellular DNA,19 which improves clearance of the sputum from the lungs. Dornase alfa significantly improves lung function, reduces pulmonary exacerbations and is used in patients with CF since 1992. The recommended dosing regimen is once daily inhalation of 2.5 mg dornase alfa, although some patients benefit from twice daily inhalation.20
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 3
|
Chapter Despite the use of inhaled antibiotics and dornase alfa in CF, lung disease still progresses. It has been recognized that inhalation therapy using one of the above described drugs results in high concentrations of the drugs in sputum. 15,21 However, concentrations in the small airways after nebulization are largely unknown. For antibiotics we know that concentrations above a certain threshold are required before they become effective. This threshold is called the minimal inhibitory concentration (MIC). Hence, it is well possible that in partly or completely obstructed airways concentrations remain below this thresh- old resulting in sub optimal treatment of these diseased lung areas. This is likely to occur especially in the small airways. Suboptimal concentrations might thus contribute to the progression of CF lung disease. Nebulizers and deposition Several factors are of influence on the amount of drug that is able to bypass the upper airways and can be deposited after a long journey in the small airways. Particle related factors are the shape, size and density of the particles. Smaller particles are more likely to bypass the upper airways and of being transported to and deposited in the small airways. Patient related factors include the diameter of the airways, the breathing pattern of the patient and structural abnormalities of the airways due to the disease. Children have smaller airways and higher inspiratory airflows relative to adults. These two factors lead to more central airway deposition. Clearly there are many factors that determine whether sufficient drug reaches the diseased areas of the lung where we most want it. Also the type of nebulizer is of great influence on the lung deposition. There are many different types of nebulizers, which have different mechanisms of aerosol generation. There are jet, ultrasonic and vibrating mesh nebulizers, with or without smart nebulizer technology. Traditionally, conventional jet nebulizers are used for nebulization of antibiotics. Tobramycin and colistin are registered for use with the Pari LC Plus nebulizer (Pari GmbH, Starnberg, Germany). 22 The Pari LC Plus nebulizer is a reusable breath-enhanced jet nebulizer without smart technology. A compressor provides continuous aerosol delivery. As aerosol is wasted during exhalation, this type of nebulizer is relatively inefficient. AZLI is registered for use with an electronic nebulizer, the eFlow (PARI Innovative Manu- facturers; Midlothian, USA). This is a vibrating mesh nebulizer. Mesh nebulizers are more efficient than jet nebulizers because there is virtually no loss of drug during exhalation and they have only a small residue after nebulization.With smart nebulizers, aerosol is only delivered to the patient during a pre-set fraction of the inspiration. The Akita is an example of a smart electronic system in combination with a jet nebulizer. This is a controlled-inhalation device that directs the flow and depth of each inhalation by coaching the patient in correct inhalation technique. This nebulizer
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 4
|
General introduction is known to increase deposition of aerosol in the small airways. 24 Moreover, the Akita has shown to improve efficacy of inhaled dornase alpha by 70% compared with 10- 20% for the standard jet-nebulizer.25 The I-neb is another example of a smart electronic system in combination with a mesh nebulizer. It works in a similar way as the Akita. Both smart nebulizers allow monitoring of patient adherence to treatment. 26 Overall smart nebulizers are more efficient and can potentially achieve higher lung doses compared to traditional jet nebulizers. Adherence and inhalation technique For correct use and for the inhaled drugs to be effective, the patient needs to be adher - ent to the treatment regimen. The treatment of CF lung disease is complex and takes one to two hours per day. Most of this time is spent on nebulizer therapy. Adherence to treatment decreases with the duration and complexity of treatment. Only 32% of patients with CF is fully adherent to a twice or trice daily treatment regimen of nebulized antibiotics.27,28 However, even if patients take medication daily, the delivery of drug into the lungs may fail due to an incorrect inhalation technique. 29,30 A poor inhalation technique reduces the amount of deposited drug at the site of action and thus reduces the effectiveness of medication. For this reason, much attention is given to instructions on inhalation technique by the CF nurses. Surprisingly little research has been done to evaluate the efficacy of these instructions in CF patients. For asthmatic children, it is known that technique related to inhalation therapy is poor. An incorrect inhalation technique was observed in 22-79% of these patients. 31-33 For CF this has never been studied. What happens in the home situation or what mistakes are made is not known. Aims of the study General aim We aimed to improve treatment of small airways disease in CF by improving the efficacy of current inhaled drugs. Specific aims Estimating drug concentrations • To develop a patient-specific airway model to predict concentrations of inhaled drugs throughout the bronchial tree in patients with CF. • To study the impact of structural lung changes and breathing profile on local drug concentrations in the airways of patients with CF.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 5
|
Chapter ABsTrACT Dry powder inhalers (DPI) delivering antibiotics for the suppressive treatment of Pseu- domonas aeruginosa in cystic fibrosis patients were developed recently and are now increasingly replacing time-consuming nebuliser therapy. Noninferiority studies have shown that the efficacy of inhaled tobramycin delivered by DPI was similar to that of wet nebulisation. However, there are many differences between inhaled antibiotic therapy delivered by DPI and by nebuliser. The question is whether and to what extent inhalation technique and other patient-related factors affect the efficacy of antibiotics delivered by DPI compared with nebulisers. Health professionals should be aware of the differences between dry and wet aerosols, and of patient-related factors that can influence efficacy, in order to personalise treatment, to give appropriate instructions to patients and to better understand the response to the treatment after switching. In this review, key issues of aerosol therapy are discussed in relation to inhaled antibi- otic therapy with the aim of optimising the use of both nebulised and DPI antibiotics by the patients. An example of these issues is the relationship between airway generation, structural lung changes and local concentrations of the inhaled antibiotics. The pros and cons of dry and wet modes of delivery for inhaled antibiotics are discussed.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 6
|
Inhaled antibiotics: dry or wet? InTroduCTIon Cystic fibrosis (CF) lung disease results in abnormal secretions in the lung that foster infec- tion and inflammation, even early in life.34,35 The vicious cycle of infection, inflammation and thick pulmonary secretions leads to early structural lung damage and to abnormal pulmonary function tests. In addition to bronchiectasis, small airways play an important role in early CF lung disease. 5 In advanced lung disease, the geometric changes related to small airways disease are significantly more severe relative to large airway changes. Progressive bronchiectasis and small airways disease eventually lead to end-stage lung disease.36 It has long been recognised that pulmonary infection, particularly by Pseu- domonas aeruginosa, is associated with progressive structural lung damage. 35 For this reason, after showing that eradication therapy using nebulised antibiotics was effective in preventing chronic infection by P. aeruginosa,14,37,38 this has become standard of treat- ment.39,40 For those patients who develop chronic P. aeruginosa infection, maintenance treatment using nebulised antibiotics has become the standard of treatment to suppress this microorganism chronically.39-Nebulised antibiotics against P. aeruginosa were developed as an alternative for intravenous therapy to deliver high concentrations of the antimicrobial agent directly to the site of infection, with the dual aims of improving efficacy and reducing toxicity. Nebulised tobramycin, colistin, and aztreonam lysine are the most commonly used nebu- lised antibiotics for this indication. Of these inhaled antibiotics tobramycin inhalation solution (TIS) has been studied most extensively. TIS should be used in a "1 month on and 1 month off" treatment cycle. Maintenance treatment with TIS has been shown to reduce exacerbations, improve lung function and improve quality of life.41,43 In addition, early treatment using inhaled nebulised tobramycin against P. aeruginosa given at the time of first isolation prevents chronic infection in about two-thirds of patients. 14,44,45 Similar efficacy was recently observed in a study with nebulised aztreonam lysine.46 Fur- thermore, treatment with nebulised aztreonam lysine has been shown to be effective in delaying the need for inhaled or intravenous anti-P. aeruginosa antibiotics for pulmonary exacerbations in CF patients, and in an improved quality of life. 47 Nebulised colistin is generally used as continuous treatment to suppress P. aeruginosa growth.18,Until recently antibiotic maintenance treatment for chronic P. aeruginosa infection could only be delivered by nebulisers. Unfortunately, nebulisers have many disadvan- tages such as the need for rigorous cleaning after each use to reduce the risk for contami- nation, they are relatively bulky to carry around and nebulisation time, particularly for older systems can be lengthy. More recently, a tobramycin inhalation powder (TIP) inhaler was developed as a more patient-friendly alternative to TIS. 49 In general, dry powder inhalers (DPIs) allow fast delivery, are more portable, require minimal cleaning and are disposable, reducing the risk of contamination. Similarly, a DPI for colistin (Colobreathe;
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 7
|
Chapter Forest Laboratories Inc., New York, NY, USA) has also recently been developed.18,50 Other antibiotics that are in development as DPI formulations are ciprofloxacin, levofloxacin, vancomycin and clarithromycin. 51-55 Regulatory studies of TIP and Colobreathe have shown non-inferiority relative to the nebulised solutions. However, even though these regulatory studies of TIP and Colobreathe showed equal efficacy of the DPI compared with the nebulised inhaled formulation, one should take into consideration that there are many patient- and device-related differences between the two inhalation modalities that might affect the efficacy of treatment. It is unlikely that the efficacy of antibiotics delivered by DPI is equivalent to that of wet nebulisation for all patients. Health professionals prescribing inhaled antibiotics for the treatment of chronic P. aeruginosa in CF or in other patients groups such as non-CF bronchiectasis56,57 should be well aware of differences in administration between DPIs and nebulisers to allow them to identify patients who might benefit from switching from nebuliser to DPI treatment, and to enable them to give appropriate instructions to patients. The aim of this review is to discuss key issues related to inhaled antibiotic therapy to optimise the effectiveness of both nebulised and DPI antibiotics. AErosol PArTIClEs And dEPosITIon For efficient aerosol treatment of both central and small airways, it is important to con - sider a number of factors that determine whether a sufficient fraction of the inhaled particles are able to bypass the upper airways and to be deposited onto the target area, namely the large and small airways. These factors can be divided into particle-related factors and patient-related factors. The aerodynamic behaviour of a particle depends on shape, size and density of the particles. The size distribution of an aerosol is usually described as the mass median aerodynamic diameter (MMAD), which refers to the droplet diameter above and below which 50% of the mass of drug is contained. In general aerosol particles smaller than 5 μm are thought to be respirable. However, particles with a MMAD between 2 and 5 μm have a lower probability of bypassing the upper airways and of being transported to and deposited in the small airways relative to 1-2 μm particles ( fig. 1). Unfortunately, small particles carry little drug. In addition to the geometric size of the particle, the particle density determines transport velocity and deposition probability. Spheres that have the same transport velocity exhibit the same aerodynamic behaviour and have similar deposition patterns in the lung. This means that particles that are large geometrically and are porous ( i.e. have a low density) will behave aerodynamically like particles that are small geometrically and are nonporous (i.e. have a high density). This effect of density
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 8
|
Inhaled antibiotics: dry or wet? on aerodynamic diameter is being used in the development of DPI drug formulations containing dry porous particles. Patient-related determinants of lung deposition and distribution within the airways in- clude, firstly, the diameter of the large airways. Children have smaller airways and higher inspiratory airflows relative to adults, both of which facilitate central airway deposition (fig. 1).58 The second patient-related factor determining particle deposition is the quality of the inhalation manoeuvre. This quality depends on age, physical capability, disease severity and the cognitive ability of the patient to perform specific inhalation manoeu - vres. It is well recognised that even well-trained and capable patients might vary their inhalation technique considerably from day to day. A high inspiratory flow rate will result in more turbulence in the central airways and, therefore, result in an increased deposition of drug in the upper airways. 59 A slow inhalation manoeuvre, however, will result in less turbulence in central airways and, therefore, in a higher probability of aerosol particles bypassing the central large airways. Hence, ideally, an aerosol should be inhaled using a slow and deep inhalation so even large particles containing a high drug mass have a higher probability of bypassing the central large airways and making it all the way down into the diseased small airways. The third patient-related factor that determines particle deposition is the presence of structural abnormalities of the airways and/or mucus in the airways, which both can result in disturbance of the airflow pattern and thus in increased Figure 1. Schematic view of the bronchial tree showing the relation between airway size, flow veloc- ity, and deposition of three differently sized aerosol particles (1, 3 and 5 µm). a) In the healthy lung, the 5-μm particle has the highest probability of being deposited onto the mucosa of the central airways due to inertial impaction. b) In the healthy lung of a child, the airways are narrower and flow velocities of the inhaled particles are higher. As a result, the 3- and 5- μm particle are deposited on the central airway mucosa. c) In the diseased lung of a child, the airways are thickened due to mucosal swelling by inflammation and mucus. As a result the cross-sectional diameter of the central airways is even smaller relative to the healthy lung (b); in addition, mucus depositions cause more turbulence of the inhaled air. As a result the 1-, 3- and 5- μm particles are deposited on the central airway mucosa. Reproduced and modified from Tiddens. Ital J. Pediatr 2003 (vol 29:39-43) with per- mission from the publisher.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 9
|
Chapter deposition at the sites of obstruction. 59-61 The fourth patient-related factor is the abil- ity of the lung to expand. Recent modelling studies showed that lobes with substantial structural damage received less inhaled antibiotic.61 It is likely that structural abnormali- ties such as fibrosis in CF lungs have a negative impact on lung expansion (fig. 2 ).62 As a result, there is a preferential airflow to the healthier regions of the lung. Hence, to select the most appropriate inhalation device for a patient, we should not only take aerosol characteristics of the inhaled drug into account but also age, inhalation flow pattern related to the device and the severity of CF lung disease. loCAl ConCEnTrATIons oF AnTIBIoTICs It is generally believed that inhaled antibiotics are so effective because of the high spu- tum concentrations that were observed in the pivotal studies.41,42 However, this concept is probably overly simplistic for a number of reasons. Firstly, the high concentrations mea- sured in sputum are most likely to reflect drug primarily deposited in the large airways. As discussed, high central airway deposition can be the result of high turbulent flows leading to upper and central airway deposition, especially of larger inhaled particles, due to inertial impaction ( fig. 3). Secondly, high concentrations in the central airways mean that less drug is available for the remainder of the bronchial tree, especially for the small Figure 2. Schematic representation of the distribution of aerosol particles throughout the lung for the normal lung and the diseased lung. a) Distribution of aerosol particles for a patient with normal lung lobes while the patient is inhaling the antibiotic using tidal volume breathing. Note that there is homogeneous distribution of the antibiotic between lung segments, and deposition is higher in the central airways relative to the small airways and parenchyma. b) Same patient as in (a) inhal- ing deeply. Note that more drug reaches the small airways and parenchyma. Furthermore, there is equal expansion of the two lung lobes. c) The patient has considerable lung disease in one lobe, and is inhaling quickly and deeply. The diseased lobe has a higher airway resistance and reduced compliance relative to the healthier parts of the lung. The expansion of the diseased lobe is slower relative to the healthier parts. As a result, there is preferential flow to the healthier lobe. In addition, the partial obstruction of the airway to the diseased lung lobe causes a turbulent flow pattern and increased aerosol deposition. The overall result is that the healthy lung lobe receives more antibi- otic relative to the diseased lobe.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 10
|
Inhaled antibiotics: dry or wet? airways. It should be kept in mind that drug bypassing the central airways is distributed over the vast surface area of the small airways. It is estimated that the total surface area of the small airways in adolescents is in the order of 1.2-1.3 m 2. For each consecutive airway generation, concentrations will be lower as the total surface area of the airway surface increases exponentially. In addition, the velocity of airflow drops from central airways towards the small airways. This changes the principal mechanism of deposition by inertial impaction in more central large airways to sedimentation by gravitational forces in the smaller airways.63 Thirdly, as discussed in the previous paragraph, diseased areas will receive a lower dose of the drug.61 Downstream to the site of obstruction, less drug will be available for deposition. Fourthly, airflow is preferentially directed towards the healthier regions of the lung (fig. 2). Taking all these factors into account, it is highly likely that a wide range of sputum concentrations exist throughout the lung, and that concentrations in the small airways are likely to be low and might even fall below thera- peutic levels depending of the mass of drug inhaled and on the aerosol characteristics.61 Whether the use of a DPI or nebuliser for the inhalation of a specific antibiotic results in differences in the surface areas with very high and subinhibitory levels is unknown but should be considered. What could the consequences of the aforementioned issues be for clinical practice? Firstly, it is unclear whether high antibiotic concentrations in central airways well above the minimal inhibitory concentration (MIC) translate to more effective killing of P. aerugi- nosa. If not, this should be considered a waste of drug that could better be delivered to Figure 3. Schematic view of the total cross-sectional area of the bronchial tree. On the left, the narrow diameter of the trachea is shown. The total cross-sectional diameter and surface area of all small airways is substantially larger relative to the trachea. Note that the inflammatory thickening of the airway wall in the small airways is more severe in the small airways relative to the large airways. Furthermore, note that the density of deposited particles is larger and, thus, antibiotic concentration is higher in the central airways relative to the small airways. In addition, large and small particles are primarily deposited in the central large airway. In the small airways, more small particles can be observed and only few large particles. B) In the same bronchial tree as in (a), Pseudomonas ae- ruginosa are distributed throughout the bronchial tree, assuming equal density. Note that antibiotic concentration in the small airways is lower than in central airways; as a result, concentrations might fall below the minimal inhibitory concentration and be insufficient to adequately treat P. aeruginosa.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 11
|
Chapter the small airways and to more diseased regions with possibly subinhibitory concentra- tions. Secondly, low/subinhibitory concentrations in the more diseased and obstructed areas are not effective, and can lead to on-going infection in these regions. Hence, even when spirometry outcome measures improve during a cycle of inhaled antibiotic treat - ment, it is likely that the diseased areas of the lung remain undertreated. This could explain the effectiveness of treatment using intravenous rather than inhaled antibiotics, which may result in more effective antibiotic concentrations in diseased areas. Thirdly, subinhibitory concentrations can lead to the development of resistance. The develop- ment of resistance in patients on chronic treatment with inhaled antibiotics has been well studied. Indeed, for TIS, a small increase in resistance has been observed. However, this development of resistance did not correlate with any efficacy parameter, 64 but it is doubtful that the resistance of expectorated sputum gives an accurate reflection of the distribution of resistance throughout the lung. It has been well recognised that the distribution of microorganisms in the CF lung is inhomogeneous. Hence, it is not surpris- ing that, even for a single species, a wide distribution of MICs can exist.65,Clearly, for inhaled antibiotics, it is important to obtain sufficiently high concentra - tions throughout the bronchial tree. In the next sections, the basics and pro and cons of nebulisers and DPIs for the delivery of inhaled antibiotics are discussed for a better understanding of the relationship between the delivery device and airway concentra- tions of inhaled antibiotics. WET nEBulIsErs: ThE BAsICs Nebulisers convert liquid medication into a mist and can be used to deliver a wide range of drug formulations for inhalation. There are different types of nebulisers, which have different mechanisms of aerosol generation. The differences in delivered aerosol between nebuliser systems currently available are significant.67 There are jet, ultra-sonic and vibrating mesh nebulisers, with or without smart nebuliser technology. The most frequently used systems to nebulise antibiotics are jet nebulisers. Jet nebulisers consist of a compressor that sucks air from the environment through an air filter and generates airflow through the nebuliser containing a Venturi tube. In the Venturi, the airflow is mixed with the fluid and a primary aerosol is formed. A baffle in the nebuliser further disintegrates the droplets into smaller aerosol particles. Most of the generated aerosol particles will fall back in the medication cup and will be re-nebulised. The patient inhales the aerosol while tidal breathing from a reservoir through a mouthpiece or face-mask. For children, a good face mask design is important to maximise efficiency. 68,69 A child old enough to inhale through a mouthpiece should do so, as the efficiency of aerosol delivery can be doubled relative to inhalation by face mask.70
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 12
|
Inhaled antibiotics: dry or wet? The recently introduced mesh nebulisers use either a vibrating or fixed membrane with a piezoelectric element with microscopic holes to generate an aerosol. Vibrating mesh devices have a number of advantages over jet nebuliser systems. They are very efficient because there is almost no loss during exhalation and the residue after nebu - lisation is usually <0.3 mL, whereas in a jet nebuliser, 1-1.5 mL is left. Mesh nebulisers are silent and are generally portable, as they operate as effectively when using batteries as when using mains power. Lung deposition of mesh nebulisers is more efficient than conventional nebulisers, varying between 30% and 80% of the loading dose, depending on the device.For smart nebulisers, the breathing pattern can be set in such a way that aerosol is only delivered during a pre-set fraction of the inspiration. In addition, for some nebulisers, the depth of the inhalation manoeuvre and the flow rate of the inhalation can be set. The deeper the inhalation, the shorter the nebulisation time. Almost no drug is lost during exhalation. Smart nebulisation systems can incorporate either jet or mesh nebulisers but the release of drug is controlled electronically, rather than allowing continuous release of drug. The I-neb (Philips Respironics, Parsippany, NJ, USA) is an example of such a system using a mesh nebuliser, while the Akita (Activaero, Munich, Germany) is an example of a smart system using a jet nebuliser. Smart nebulisers are substantially more efficient and may achieve lung deposition of 60-80% of the loading dose (using mesh technology), 71 compared to 5-15% with traditional jet nebulisers.22 Furthermore, more efficient deposition of aerosol in the small airways can be achieved, which may result in more effective treatment of small airway obstruction.Although there are many advantages, vibrating mesh and smart nebulisers are still not extensively tested in children and there is little clinical information available. Lung deposition is improved but evidence for dosage recommendations is either lacking or based on in vitro or adult data. New-generation nebulisers for inhaled antibiotics can be used as long as efficacy and toxicity data are available, especially when using potentially toxic antibiotics such as inhaled tobramycin. nEBulIsErs For InhAlEd AnTIBIoTICs; Pros And Cons Inhaled antibiotics are registered for use with a specific nebuliser-compressor combina- tion. Phase III studies with TIS were performed with the LC Plus nebuliser (Pari GmbH, Munich, Germany) with either the Pari Turbo Boy or PulmoAide compressor (DeVilbiss, Mannheim, Germany).41 The nebulisation of colistin is registered for use with the Pari LC Plus nebuliser and with the I-neb nebuliser in the UK. However, no proper Phase III registration studies were conducted for nebulised colistin.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 13
|
Chapter For the nebulisation of aztreonam for treatment of chronic P. aeruginosa airway infec- tion, the Pari eFlow mesh nebuliser has been registered.The use of nebulisers has some well-recognised advantages. 1) They are a platform to deliver drugs that are only available as fluids. 2) They can be used for all ages from infancy into adulthood. 3) The nebulised drug can be inhaled while the patient is breath- ing tidally. Hence, no specific inhalation manoeuvre is required. 4) Over the last few decades, smart nebulisers such as the I-neb and the Akita have been developed that allow electronic data logging. This supplies objective information for patient and CF team on treatment adherence.25,28 5) Smart nebulisers such as the Akita allow the control of inhalation competence. Hence, the patient has to follow a pre-set inhalation profile, optimising treatment efficacy and efficiency. 6) Smart nebulisers can be set to efficiently target the small airways, which play an important role in CF lung disease. 5,72 To do so, a personalised inhalation profile can be programmed onto a smart card in the nebuliser. Improved delivery of dornase alfa to the small airways using the Akita nebuliser resulted in substantial improvement of small airways patency. Similar approaches could theoreti- cally be of benefit to improve efficiency and efficacy of other nebulised drugs such as hypertonic saline and antibiotics. There are some important disadvantages related to nebuliser therapy. 1) It is time- consuming. For patients on maintenance treatment with dornase alpha and inhaled anti- biotics, nebuliser therapy can take up 2 h per day. 73 This time is needed for preparation of the nebuliser, nebulisation of the drug and cleaning the nebuliser after its use. For colistin, up to 10 minutes extra time is needed to dilute the appropriate dose in water for injection to obtain an isotonic solution. This is not the case for TIS or aztreonam lysine, which are readily available in a unit-dose vial. 2) The use of nebulisers includes the risk of contamination when not properly cleaned. 74 2) Nebulisers are bulky and less portable than other devices. 4) Nebulisers require regular maintenance. Over time, the air filter of jet nebulisers gets polluted with dust particles and, thus, should be replaced at regular intervals. Similarly, 'lifelong' nebulisers suffer from wear and tear, and thus require replacement once to twice a year. Furthermore, the compressor output of jet nebulisers should be periodically examined as per the manufacturer's instructions. For the Pari eFlow mesh nebuliser used for aztreonam for inhalation solution (AZLI; Cayston (Gilead Sciences Inc., Forest City, CA, USA)) therapy, maintenance issues are different. Occlusion of the holes can occur, 75 which prolongs treatment time. Hence, they require careful cleaning after each use and frequent replacement of the mesh to prevent build- up of deposit and blockage of the apertures. 75 Therefore, a new mesh is delivered with each monthly package of aztreonam. In addition, the mesh should be replaced if nebu- lisation time exceeds 5 min. Overall, a great need was felt by the CF community and the pharmaceutical industry to develop more time-efficient, less cumbersome alternatives for nebulized antibiotic therapy.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 14
|
Inhaled antibiotics: dry or wet? dPIs: ThE BAsICs In DPIs for antibiotics, the drug is present as a dry powder formulation in a capsule. The loading dose of the capsules for antibiotics is in the 20-150-mg range and, therefore, substantially higher than that of antiasthma drugs, which is mostly in the 50-200-μg range. For this reason, multiple capsules can be needed to inhale a sufficient mass of the antibiotic into the lungs. The technical properties of the dry powder formulation combined with the properties of the inhaler determine how the powder can best be inhaled. Both the mass and the aerosol characteristics of the released aerosol depend on the inhaled volume and on the inspiratory flow profile generated by the patient. TIP is formulated using PulmoSphere technology (Novartis AG, Basel, Switzerland). This is a spray-drying technique, which generates relatively large porous particles that disperse very easily. For each treatment, four capsules of 28 mg TIP need to be inhaled through a low-resistance inhaler to get a lung dose equivalent to that of 300 mg nebulised TIS. A low flow of 30 L/min is sufficient to release the drug from the capsule and to disperse the aerosol particles. TIP has an MMAD <4 μm and a grain size distribution of 1.7-2.7 μm.76 The low-resistance inhaler allows the patient to generate a wide range of inspiratory flows.77 In a controlled laboratory setting, it was shown that almost all CF-patients of 6 years and older were able to generate flows of ≥30 L/min. However, some patients obtained inspiratory flows as high as 170 L/min through a low-resistance inhaler. Unfortunately, as discussed, high inspiratory flows increase oropharyngeal deposition and can reduce lung deposition, especially in the small airways. To empty all drug from the TIP capsule, an inspiratory volume of 1 L is sufficient to release all dry powder.77 Most patients of 6 years and older were able to inhale a volume of ≥1 L.77 However, to ensure that all drug is released from the capsule, it is recommended to repeat the inhalation manoeuvre twice for each capsule. Colobreathe is formulated as micronised particles that, in general, do not disperse very easily. For each administration, one 125-mg capsule needs to be inhaled through a low-resistance inhaler to get a lung dose equivalent to that of 160 mg nebulised co- listin. It is claimed that an inspiratory flow of 30 L/min through the inhaler is required to disperse the micronised drug optimally into respirable aerosol particles. 47 However, currently, there are no published data available describing the aerosol characteristics of the colistin DPI. To ensure that all drug is released from the capsule, it is recommended to repeat the inhalation manoeuvre twice for each capsule. Clearly, the inspiratory flow and volume when inhaling an antibiotic from a DPI are important determinants of the deposition pattern and efficacy. Surprisingly, the optimal inhalation profiles of TIP and Colobreathe® have not been clearly defined to date; this should be further investigated to optimise this form of inhaled antibiotic treatment. Next, patients should be trained to use the optimal inhalation technique with training aids and this technique should be regularly evaluated.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 15
|
Chapter DPIS: PrOS AND CONS DPIs have major advantages over nebulisers. 1) Administration is quick. For TIP , four capsules can be easily inhaled in 5 min. 2) DPIs do not require extensive cleaning after use. 3) Maintenance of the DPI is not required. 4) They are easy to carry around. 5) The capsules are packaged in sealed blisters and do not require refrigeration. Overall, this method of delivery is more convenient and sterile, which is an important consideration for CF patients who are highly susceptible to lung infections. The most important disadvantage of a DPI inhaler is that the aerosol characteristics of the inhaled drug and, therefore, of the lung deposition can be highly dependent on the inhalation profile generated by the patient through the DPI. To our knowledge, no field studies have been published that observed how CF patients operate these DPIs in daily life. The DPIs for TIP and colistin are low-resistance inhalers. Hence, when a patient inhales forcefully, very high inspiratory flows can be obtained, resulting in a high oropharyngeal and central airway deposition. 59 In addition, this can result in cough. In the TIP versus TIS study, cough was reported as adverse event in 48% of the subjects on TIP versus 31% in the patients on TIS. 49 The optimal inhalation profile is formulation dependent. Hence, for the spray-dried, hollow porous particles of TIP , a slow and deep inhalation might be sufficient to disperse the drug while reducing upper airway deposition and improving deposition into the small airways. The optimal inhalation for the micronised colistin formulation is difficult to predict. It is likely that the dispersion of this formulation is highly flow dependent. A very high inspiratory flow might be needed to generate a Table 1. Inhaler devices for currently available inhaled medication in cystic fibrosis Inhaled drug Nebuliser DPI pMDI hypertonic saline (7%) + - - Mannitol - + - dornase alpha + - - Bronchodilators + + + Inhaled corticosteroids + + + Tobramycin + + - Colistin + + - Aztreonam + - - AmBisome + - - liposomal Amikacine + - - Ciprofloxacin - + - Vancomycin - + - Clarithromycin - + - AmBisome is manufactured by Gilead Sciences, Uxbridge, UK. DPI: dry powder inhaler; pMDI: pressurised, metered-dose inhaler.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 16
|
Inhaled antibiotics: dry or wet? sufficiently large fraction of small particles to treat the small airways effectively. For dry powder antibiotic formulations that are still in development, it will be important to edu- cate prescribing physicians and patients to understand the key characteristics of each formulation (and device) and to be aware of the optimal inhalation profile required for that formulation. Another factor to consider is that when DPIs are prescribed to a patient for antibiotics and other medications that require different inhalation patterns ( table 1), this is likely to result in confusion and erroneous use. Finally, we should investigate whether patients can be trained consistently "not inhale too fast" or if the devices can be modified to ensure that patients inhale within the correct range of inspiratory flows (e.g. by increasing the device resistance or the use of visual/auditory aids). dry or WET? Taking into account the pros and cons of nebulisers and DPIs for maintenance antibiotic treatment, it is clear that DPIs are more convenient for patients and less conspicuous to use, and from this perspective, the device of preference for patients. However, the most important reason for the physician to select one or the other should be primarily based on effectiveness. When selecting the potentially most effective inhalation device for/ with the patient, several considerations should be taken into account (table 2). In the development program of TIP , it was designed to be equally effective as TIS. To accomplish Table 2. Considerations when prescribing inhaled antibiotics The concentration gradient of an inhaled antibiotic goes progressively down from central airways towards the small airways There is a preferential flow of inhaled antibiotic towards the more healthy regions of the lung The more diseased the lung, the more inhomogeneous the deposition pattern and the more regions will be suboptimally treated Subinhibitory concentrations for inhaled antibiotics are likely to occur in advanced disease Inhalation of an antibiotic by DPI is faster and cleaner relative to nebulised antibiotics The aerosol characteristics of an inhaled antibiotic by DPI depend on formulation, device and inhalation manoeuvre Each antibiotic inhaled by DPI has a device- and formulation-specific optimal inhalation profile The patient (and parents in the case of children) should both be aware of the optimal inhalation profile The efficacy of inhaled antibiotic therapy is determined by adherence and inhalation competence Inhalation technique should be repeatedly evaluated and patients (parents) repeatedly trained In case of a suboptimal therapeutic treatment result, check and recheck inhalation competence For patients using a DPI but who cannot reproducibly generate the optimal inhalation profile, consider switching back to a nebuliser or to a smart nebuliser that guides the patient in optimising the inhalation manoeuvre DPI: dry powder inhaler.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 17
|
Inhaled antibiotics: dry or wet? disease; are all treated with the same regimen. It might well be that for more advanced disease, a higher dose is needed to cover all airway generations with antibiotic con- centrations above MIC. For TIP and TIS, a once daily double dose might result in larger areas with concentrations above MIC and may therefore be more efficacious than current twice-daily administration. In addition, because of the difference in mechanism of action of various antibiotic classes, the frequency of administration should be optimised for each class. Advanced mathematical modelling can help us to determine this relationship and to design a specific dose relative to disease severity. Furthermore, it might be pos - sible to improve the effectiveness of eradication therapy by increasing the dose in those patients with more advanced disease and in whom primary eradication therapy fails. Current DPI devices do not control the inhalation flow by the patient. Hence, there is an opportunity to optimise the inhalation manoeuvre of recently developed DPIs based on their characteristics. Training aids will be needed to facilitate this training. In the future, smart DPIs might be developed that guide the patient through the optimal inhalation manoeuvre. ConClusIon Inhaled antibiotics are of key importance in the treatment of CF-related lung disease. Care should be taken to ensure that the small airways are efficiently targeted, even in diseased regions of the lung. Whether this is the case depends on many factors such as age, inhalation manoeuvre, severity of structural lung disease and other factors. Nebulis- ers are important especially for those inhaled antibiotics that are only available as a fluid. The use of nebulisers requires that technical maintenance is well organised. When possible, DPIs should be used to reduce the treatment burden. CF caregivers and patients should be aware that there are major differences between the inhalation manoeuvre of a nebulised antibiotic and a DPI. Aerosol deposition by DPIs can vary widely in relation to the inhalation manoeuvre. Hence, switching a patient from a nebuliser to a DPI requires careful instruction of the optimal inhalation manoeuvre for that specific antibiotic. The optimal inhalation manoeuvre should be clearly defined by the pharmaceutical industry. All aspects of inhaled antibiotic therapy should be carefully and frequently evaluated with the patient in the starting phase and, when used routinely, at least once a year. Alternative dose regimens for inhaled antibiotics need to be further investigated.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 18
|
Chapter ABsTrACT Background: Chronic airway infections are an important factor in progressive lung disease in patients with cystic fibrosis (CF). These infections are most often treated with inhaled antibiotics of which deposition patterns have been extensively studied. However, the journey of aerosol particles does not end after deposition within the bronchial tree, but continues through thick mucus layers and biofilm generated by bacteria. Objectives: To review what happens to antibiotic aerosol particles after deposition in the airways of patients with CF and how local conditions affect its clinical efficacy. methods: We searched Embase, Medline, Web-of-Science, Scopus, Cochrane, PubMed publisher and Google Scholar databases from inception to September 2015. Original studies describing the effect of CF sputum or bacterial factors on antibiotic efficacy and liposomal formulations or co-medication to increase efficacy were included. Two authors independently assessed the study eligibility of the selected publications. results & conclusions: 2669 articles were screened of which 35 met the inclusion criteria for this review. Based on these articles, which mainly consisted of in vitro studies, we conclude that the clinical efficacy of inhaled antibiotics can be reduced by many fac- tors after deposition in the airways. Aminoglycosides were the most extensively studied antibiotic and are adversely affected by molecules within CF mucus and the alginate layer surrounding Pseudomonas aeruginosa.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 19
|
Chapter Articles were selected based on the following inclusion criteria: (i) effect of CF sputum or (ii) bacterial factors on antibiotic efficacy; (iii) local efficacy determined by antibiotic concentration or number of molecules; (iv) liposomal formulations or co-medication to increase efficacy; (v) original research. The exclusion criteria are as follows: (i) article not in English or Dutch; (ii) data solely on clinical efficacy of inhaled antibiotics; (iii) no ab - stract or full text available; (iv) inhaled non-antibiotic drugs; (v) antibiotic combinations; (vi) nanoparticles to increase efficacy (vii) review/overview. Both reviewers assessed the full text of each selected article to ensure that they met the eligibility criteria, developed to critically appraise the selected publications. These criteria were based on the Grading Recommendations Assessment Development and Evaluation (GRADE) criteria by the GRADE working group and the scoring methods of descriptive studies by Slim et al.82,83 The set of eligibility criteria was more extensive for comparative studies than for non-comparative studies. Using these criteria the relevance and validity of the selected articles were scored independently by the two reviewers. Each criterion received 0, 1 or 2 points (online supplementary Table S1) and the total Table 1 – Search terms Database searches Embase ('cystic fibrosis'/de OR 'lung fibrosis'/exp OR (((cyst* OR lung OR pulmonar*) NEAR/3 fibro*) OR fibrocyst* OR mucoviscid* OR cf)) AND (((('antibiotic agent'/exp OR 'antibiotic therapy'/de OR 'antiinfective agent'/de OR 'antibiotic sensitivity'/exp OR 'antimicrobial activity'/exp OR levofloxacin/de OR (antibiotic* OR antimicrob* OR antibact* OR (anti NEXT/1 (biotic* OR microb* OR bact*)) OR tobramycin* OR colisti* OR colisiti* OR colomycin* OR colymycin* OR (coly NEXT/1 mycin*) OR tadim OR aztreonam OR aminoglycoside* OR amikacin* OR levofloxacin* OR 'mp 376' OR azithromycin* OR vancomycin* OR gentamicin OR bactericid*):ab,ti) AND (inhalation/de OR 'oral spray'/de OR 'inhalational drug administration'/de OR 'nebulization'/de OR inhaler/exp OR nebulizer/exp OR aerosol/de OR powder/exp OR (inhal* OR vapor* OR vapour* OR aerosol* OR spray* OR mist OR atomi* OR nebuli* OR compressor* OR powder* OR dry OR dried OR jet OR ultraso*):ab,ti)) OR (gernebcin OR tobi OR tsi OR bramitob OR cayston OR azli OR (liposom* NEAR/3 amikacin*)):ab,ti) AND (pharmacodynamics/exp OR pharmacokinetics/exp OR 'drug efficacy'/de OR 'concentration (parameters)'/exp OR 'concentration response'/de OR 'drug sputum level'/de OR 'sputum analysis'/de OR clearance/exp OR (pharmacodynam* OR pharmacokinet* OR effectiv* OR efficien* OR efficac* OR concentrat* OR sputum* OR mucocilliar* OR mucus OR ((lining OR surface) NEAR/3 (fluid* OR liquid*)) OR clearance* OR 'half life'):ab,ti)) Medline Modelled search strategy designed for Embase Web-of-science Modelled search strategy designed for Embase Scopus Modelled search strategy designed for Embase Cochrane Modelled search strategy designed for Embase Pubmed Publisher Modelled search strategy designed for Embase Google scholar Modelled search strategy designed for Embase
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 20
|
Chapter The following antibiotic classes were studied: aminoglycosides (n=28 studies), β-lactam antibiotics (n=12), fluoroquinolones (n=2), tetracyclines (n=1) and other anti - biotics (n=9). Table 2 – Impact per antibiotic class Factor Antibiotic class Impact CF mucus in general Aminoglycosides • Reduces diffusion• Strong binding96,98,β-lactam AB • Reduces diffusion• Negligible binding94,Fluoroquinolones • Reduces diffusion90,Other (Polymyxin B) • Strong bindingMucin Aminoglycosides • Strong binding94,• Reduces efficacy of liposomal forms more strongly than free formsβ-lactam AB • Reduces diffusion• No bindingDNA Aminoglycosides • Strong binding93, 94, 97, β-lactam AB • Reduces diffusion• No bindingOther (Polymyxin B) • No bindingNucleic acids Aminoglycosides • Strong binding94, Bacterial endotoxins (lPs,lTA) Aminoglycosides • Strong bindingOther (Polymyxin B) • BindingAnaerobic conditions in mucus Aminoglycosides • Reduces efficacy80, 107, β-lactam AB • Reduces efficacy80, 107, Macrolides • Reduces efficacyFluoroquinolones • Reduces efficacy,80, 108 remain bactericidalTetracyclines • Reduces efficacyOther • Cotrimoxazol: Reduces efficacy• Chloramphenicol: Reduces efficacy• Colistin: Reduces efficacy,80 Increases efficacySalt content in mucus Aminoglycosides • Reduces efficacy (magnesium, calcium, nitrate)108, • Decreases binding to alginate (sodium chloride, potassium, calcium, phosphate)112, β-lactam AB • Carbenicillin: nitrate increases efficacy if aerobic, but effect abolished if anaerobic• Ceftazidime: nitrate no effectFluoroquinolones • Nitrate decreases susceptibility of organisms to ABTetracyclines • Nitrate no effect108
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 21
|
Systematic review: fate of inhaled antibiotics Table 2 – Impact per antibiotic class (continued) Factor Antibiotic class Impact Other (Chloramphenicol) • Nitrate no effectAlginate Aminoglycosides • Reduces diffusion81, 111, 112, • Binding81, 112, 113, β-lactam AB • Reduces diffusionTetracyclines • Reduces diffusionOther ( Polymyxin B) • Reduces diffusion• Bindingliposomal formulation Aminoglycosides • Increases efficacy,101-104 decreases efficacy• Reduces bindingOther (Polymyxin B) • Increases efficacy• Reduces bindingdornase alfa Aminoglycosides • Increases binding,95 decreases binding• Enhances bactericidal efficacy,87 reduces bactericidal efficacyFluoroquinolones • Enhances bactericidal efficacy• Enhances diffusionMannitol Aminoglycosides • Enhances bactericidal efficacyFluoroquinolones • Enhances bactericidal efficacy• Enhances diffusionAlginate lyase Aminoglycosides • Improves diffusion• Tobramycin and amikacin: Enhances bactericidal efficacy87, • Gentamicin: Enhances bactericidal efficacy,114 no effect on bactericidal efficacyNaCl Aminoglycosides • Enhances bactericidal efficacy of tobramycin,106 antagonistic effect on tobramycinFluoroquinolones • Enhances bactericidal efficacy• No effect on diffusionOther (Colistin) • Synergistic effectGlucose Aminoglycosides • Enhances bactericidal efficacyAlX-109&AlX-009 Aminoglycosides • Enhances bactericidal efficacy116-β-lactam AB • Enhances bactericidal efficacy116-Cationic amphiphiles Aminoglycosides • Reduces binding with DNACitrate Aminoglycosides • Synergistic action with amikacin• Reduces bactericidal efficacy of tobramycinMacrolides • Synergistic action with erythromycinOther • Synergistic action with colistin• Reduces bactericidal efficacy of polymyxin BSuccinic acid Aminoglycosides • Reduces bactericidal efficacy of tobramycinOther • Synergistic action with colistin• Reduces bactericidal efficacy of polymyxin BAB = antibiotic; CF = cystic fibrosis; LPS = lipopolysaccharides; LTA = lipoteichoic acid.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 22
|
Chapter 1. d issolution of antibiotic in mucus layer (Table s4, part 1) After deposition on the mucus layer, local bioavailability of the inhaled antibiotic first depends on the solubility of the drug in the mucus. Within CF patients, the secretor and non-secretor phenotype are described: exocrine secretions between these patients differ due to the presence or absence of ABH glyconjugates.84 Tobramycin was shown to dissolve faster in mucus of secretors relative to non-secretors. However, the question is if this is clinically relevant as this difference was only by 5 minutes.84,2. d iffusion through mucus layer (Table s4, part 2) Mucus primarily consists of mucin (glycoprotein) but also contains proteins, DNA, lipids and cellular debris. After dissolution, the drug needs to diffuse through the mucus layer to reach the bacteria, whereby the viscosity and elasticity of CF mucus is increased due to various factors. Mucin molecules in CF mucus are very long, extensively branched and have been shown to interact with other macromolecules in mucus secretions, in- cluding: albumin which increased its viscosity, and DNA which increased its elasticity. 86 Additionally, elevated concentrations of actin and alginate contributed to the increased viscoelasticity of CF mucus.87 Due to the long length of the mucin glycoprotein backbone and its branching, they have a high tendency to interact and obstruct drug transport in vitro,86 in particular liposomal antibiotics.For β-lactam antibiotics, the presence of mucin alone caused a 2-fold delay in the diffusion rate (1.1 to 0.5 µm/mm2/h) compared to the baseline rate determined in buffer. The addition of DNA resulted in a 10-fold delay in the rate of diffusion (0.1 µm/mm2/h).Aminoglycoside diffusion was also delayed by mucus of CF patients compared to buf- fer, but no specific numbers were reported. Furthermore, diffusion was not improved by combining gentamicin powder with the amino acid L-leucine.Combinations of ciprofloxacin dry powder with mannitol showed enhanced diffu - sion and significantly higher antibacterial activity against Pa than ciprofloxacin-NaCl or ciprofloxacin-lactose particles90 (also against Pa growing in biofilm).91 Ciprofloxacin-NaCl particles also showed higher antibacterial activity against Pa (albeit to a lesser extent), although NaCl alone had no effect on drug diffusion.90 Finally, ciprofloxacin-dornase alfa powder showed greater antibacterial activity than ciprofloxacin dry powder due to the better dissolution and diffusion abilities of ciprofloxacin. 92 Ciprofloxacin-dornase alfa powder completely diffused after 30 minutes while ciprofloxacin powder alone was not completely dissolved even after 2 hours.In summary, the diffusion rate of aminoglycosides, β-lactam antibiotics and fluoroqui- nolones through CF mucus is reduced but may be increased by coadministration with mannitol or dornase alfa.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 23
|
Systematic review: fate of inhaled antibiotics 3. m ucus binding (Table s4, part 3) Apart from delaying the diffusion of antibiotics, macromolecules present in mucus can bind to certain antibiotics and drastically reduce their efficacy, as only free drug can be active against bacteria. In particular, the efficacy of aminoglycosides (cations) is re - duced,93 due to their electrostatic interactions with anionic electrolytes.Tobramycin,94-96 amikacin97 and gentamicin96,98 have been shown to bind to mucin and free DNA. Their affinity for mucin was higher in the presence of free DNA.87,95 The latter is released from lysed leukocytes and bacteria during infections.95 In mucus collected from CF patients during exacerbations, a substantial fraction of inhaled tobramycin was bound to mucus, 94 which reduced its activity by a factor of approximately 30. 93 Binding was dependent on tobramycin concentration96 and the concentration of macromolecules,84,94 in which higher tobramycin concentrations resulted in higher concentrations of free drug. Thirty percent of tobramycin was bound at tobramycin concentrations of 5-15 μg/ ml while 15% was bound at concentrations of 25-50 μg/ml.Tobramycin (15-95%) exhibited strongest binding to mucins and DNA. 93 Up to 60% (range 1-60%) of amikacin97 and 52% (range not reported) of gentamicin was bound to mucus of patients with CF.Aminoglycosides show lower affinity to negatively charged components in mucus in an alkalotic compared to an acidic environment.98 The pH in CF airways (not measured in mucus) varies from 6.5-7.5 and seems to be constant from central to peripheral airways. In patients with pneumonia, the pH in infected bronchi was significantly lower than that in non-infected bronchi (6.48 versus 6.69).Polymyxin B and neomycin show strongly elevated minimal bactericidal concentra- tions (MBC) due to binding to CF mucus. 100,101 Polymyxin B appeared to bind to bacte- rial endotoxins within CF mucus, but not to DNA or actin filaments. 101 No binding was observed between mucus and β-lactam antibiotics94,100 and the role of mucus binding in fluoroquinolones or macrolides was not studied in the selected articles. 3.1 Methods to reduce drug-mucus interaction 3.1.1 Liposome-entrapment Most studies show that liposome-entrapment reduces antibiotic inhibition by macromol- ecules and enhances the bactericidal activity of antibiotics. Inhibition of aminoglycosides by DNA and actin filaments, and by bacterial endotoxins was reduced by 4-fold and 100- fold, respectively, when entrapped in liposomes.101 Additionally, liposomal formulations were significantly more efficacious in reducing the Pa load in rats 102 and reduced the minimal inhibitory concentrations (MICs) for Pa strains in vitro and Burkholderia ceno- cepacia strains in rats. 103,104 However, one study showed that liposomal aminoglycoside
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 24
|
Chapter activity was inhibited to a greater extent by mucins than free aminoglycosides (up to 32-fold vs up to 8-fold).3.1.2 Co-treatment with dornase alfa Conflicting results were described for the co-treatment of aminoglycosides with dornase alfa. Dornase alfa was shown to increase free tobramycin in sputum by approximately 30%94 and enhance the bactericidal activity of free and liposomal aminoglycosides. Specifically, the higher the concentration of dornase alfa, the stronger the bactericidal activity.Conversely, another study reported the decline in free tobramycin and its activity following dornase alfa treatment.95 A possible explanation given by the authors was that while dornase alfa did indeed cut the DNA into smaller strands, the charge of the strands remained unchanged. Therefore, the smaller strands were still able to bind positively charged antibiotics. 3.1.3 Cationic amphiphiles and N-acetylcysteine Cationic amphiphiles are positively charged lipid solutions that have the potential to decrease binding between DNA and tobramycin and thereby enhance its antibacterial activity. By matching the cationic amphiphiles in charge and shape, tobramycin was com- petitively displaced from DNA complexes by these agents, resulting in a 15-fold increase in tobramycin activity.Co-treatment with 1 mg/ml of the mucolytic N-acetylcysteine (NAC) did not influence the bactericidal activity of free and liposomal aminoglycosides.3.1.4 Co-treatment with Nacl Co-administration of NaCl (tested NaCl concentrations: 0.3%, 0.9%, 2.3% and 4.05%) had a synergistic effect with colistin at a concentration of 4.05% for the treatment of Pa and at all concentrations for Escherichia coli.105 Studies investigating the coadministra- tion of NaCl and tobramycin are inconclusive. Both an antagonistic effect on tobramy - cin,105 as well as improved tobramycin efficacy against young Pa biofilms106 were shown when NaCl was added. Nevertheless, in patients with CF who use one of these antibiotic formulations and inhaled saline, the timing of inhaled saline in relation to tobramycin inhalation will need to be taken into account.In summary, mucus binding appears to reduce the efficacy of aminoglycosides but not β-lactam antibiotics, and may be reduced by liposome-entrapment or coadministration of amphiphilic molecules.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 25
|
Systematic review: fate of inhaled antibiotics 4. Influence of oxygen level in mucus (Table s4, part 4) Thickened mucus layers in the lungs of CF patients contain areas of low oxygen ten- sion,107 with an anaerobic environment near the epithelial surface and higher oxygen levels at the top of the mucus layer.Aminoglycosides,80,107 β-lactam antibiotics,107,108 chloramphenicol and tetracycline 108 showed reduced efficacies under anaerobic conditions, whereas tobramycin and cip - rofloxacin were approximately twice less effective. For tobramycin, 50% of Pa isolates were killed under aerobic conditions, 30% under anaerobic conditions 80 and the log reduction dropped from 5.67±0.00 to 2.14±0.42. 108 For ciprofloxacin, the log reduction dropped from 5.05±0.31 to 2.61±0.13 under anaerobic conditions. 108 On the contrary, levofloxacin maintained its bactericidal effect under anaerobic conditions. 107 Likewise, colistin may even be more effective under anaerobic conditions. Reductions in minimum biofilm eradication concentrations (MBECs), MIC50, MIC90, MBC (2-fold, 8-fold, 4-fold and 2-fold reductions, respectively) against Pa were shown under anaerobic conditions. 109 However, another study found decreased killing of Pa isolates by colistin under anaero- bic compared to aerobic conditions (75% vs 100%).In summary, low oxygen tension reduced the efficacy of aminoglycosides, β-lactam antibiotics, tetracyclines and chloramphenicol. Colistin, however, may be more effica - cious. 5. Influence of salt content of mucus (Table s4, part 5) The antibacterial activity of certain antibiotics is highly dependent on the ionic envi- ronment.98 In CF, the ionic environment of mucus changes as a result of cell lysis; as evidenced by higher calcium levels detected in CF mucus compared to non-CF patients.86 Magnesium and calcium have a stabilizing influence on the cell walls of Pa, which is primarily driven by divalent cations, and thereby delay the effect of aminoglycosides.110 The monovalent salt sodium did not have any measurable effect on gentamicin, while the divalent magnesium salt completely shielded Pa from its activity.For E. coli, the protection by salts could be solely attributed to ionic strength and not to the type of salt. When the ionic strength was increased in vitro from 0.12 to 0.14 μ with MgCl2, NaCl or Na2SO4, gentamicin activity against E. coli ranged between 40-50%. Under anaerobic conditions, nitrate decreased the bactericidal activity of aminoglyco- sides and fluoroquinolones by half.80,108 Mucus of CF patients contains 250-350 μmol/L nitrate and concentrations can be as high as 1000 μmol/L.107 Nitrate had little effect on the efficacies of chloramphenicol, tetracycline and ceftazidime.In summary, the ionic environment is another important variable that can reduce the effectiveness of inhaled antibiotics against Pa.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 26
|
Chapter 6. Barriers generated by Pseudomonas aeruginosa 6.1 Non-mucoid Pa, Mucoid Pa and alginate formation Pa has the ability to grow under aerobic and anaerobic circumstances and exists in both a mucoid and non-mucoid formation. Chronic Pa infections are associated with more mucoid variants that produce the polysaccharide alginate and form biofilms within the lungs of patients with CF. Alginate is an important factor in the resistance of Pa against antibiotics, as it increases the colonization rate within the respiratory tract. Importantly, alginate seems to act as an ionic trapping agent for positively charged aminoglycosides and polymyxin B, thereby reducing the uptake and early bactericidal effect of antibi - otics. Additionally, alginate inhibits the non-opsonic phagocytosis of monocytes and neutrophils, thus allowing the bacteria to avoid the phagocytic immune response.87 Due to these barriers, achieving an inhibitory concentration at the surface of a mucoid colony is not sufficient to eliminate Pa. To kill the bacteria, bactericidal antibiotic concentrations need to be attained at the cell surface for a sufficient period of time.6.2 Alginate and diffusion of antibiotics (T able S4, part 5) Alginate reduces the diffusion of antibiotics in vitro and is evidenced by the fact that ami- noglycosides exhibited diffusion coefficients in alginate, which were approximately 20% of the β-lactam values (0.65 versus 3.7 x 10-6 cm2/s).81,111,112 This can be attributed to the fact that not only does the alginate itself form a physical barrier to the antibiotic, but also because the positively charged aminoglycosides (in contrast to the β-lactam antibiotics) bind to the negatively charged alginate polymers.80,81,113 This is further evidenced by the fact that 2% Pa alginate suspension completely inhibited the diffusion of gentamicin, to- bramycin and polymyxin B, whereas the diffusion of the negatively charged carbenicillin was not impeded by this suspension.114 Furthermore, diffusion rates for tobramycin were consistently lower than for gentamicin.112 The binding of alginate to antibiotic appeared to be concentration dependent81 as aminoglycosides formed precipitates with the algi- nate at a low alginate to antibiotic ratio. This phenomenon disrupted the gel structure, resulting in diffusion of aminoglycosides at a rate that was even faster than that of the β-lactams.For β-lactam antibiotics, the diffusion rate was strongly reduced by alginate as the mo- lecular weight of these antibiotics was increased. Like mucin, free DNA further reduced the diffusion of β-lactam antibiotics through alginate gels.In summary, the diffusion of both aminoglycosides and β-lactam antibiotics are reduced through the alginate layer surrounding Pa. Ultimately, this reduced diffusion contributes significantly to the difficulty in eradicating mucoid Pa from the airways of CF patients.112
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 27
|
Systematic review: fate of inhaled antibiotics 6.3 Biofilm formation Within biofilms, a subpopulation of persistent Pa cells is formed and characterised by reduced metabolic activity and tolerance to antibiotics. 106 It was shown that MBECs for three different Pa strains increased between 8 and 512 times for free aminoglycosides and 8 to 256 times for liposomal aminoglycosides when growing in biofilm.87 The efficacy of macrolides to eradicate bacteria within mature biofilms was markedly reduced relative to aminoglycosides.6.4 Therapies to overcome barriers generated by Pa Alginate lyase (AlgL) is an enzyme that can degrade alginate and facilitate the diffusion of aminoglycosides to the target bacteria.114 Co-treatment with AlgL increased bacterial susceptibility to antibiotics and phagocytosis, and reduced alginate levels. Additionally, AlgL was more effective at enhancing the activity of aminoglycosides than dornase alfa.87 This effect differed per aminoglycoside antibiotic, where AlgL treatment alone increased the bactericidal activity of tobramycin and amikacin (free and liposomal form). Likewise, the combination of dornase alfa-AlgL enhanced the bactericidal activity of tobramycin and that of free but not liposomal amikacin. 87 For gentamicin, one study showed en- hanced bactericidal activity,114 while neither AlgL nor the combination with dornase alfa demonstrated any effect on its activity in another study.An excess of iron can induce biofilm formation by Pa and is evidenced by the fact that the iron concentration in the ELF of CF patients is 400-fold higher than in non-CF patients.116 Drugs containing different combinations of lactoferrin (iron-binding glyco - protein) and hypothiocyanite (bactericidal agent; ALX-009 and ALX-109) had an additive effect on tobramycin and aztreonam in reducing both biofilm formation and disrupting established Pa biofilms. Both lactoferrin and hypothiocyanite are part of the normal in - nate immune response but their secretion by airway cells is reduced in CF.116-The addition of mannitol improved tobramycin efficacy by 99.5% against young Pa biofilms (pre-grown for 5h) and by 77% against established biofilms (pre-grown for 20h). However, mannitol had no effect on clinical strains with high resistance to tobramycin. The addition of glucose resulted in similar outcomes, albeit to a lesser extent. Similarly, NaCl required 2-fold higher osmolarities than mannitol to obtain a similar effect and had no effect on established biofilms.Finally, co-treatment with dispersion compounds (citrate and succinic acid) has been investigated to enhance biofilm eradication. Combinations of citrate with amikacin, colistin or erythromycin and succinic acid with colistin resulted in significantly enhanced killing of bacterial populations compared with control populations. However, increased bacterial viability was seen when tobramycin and polymyxin B were combined with dispersion compounds.115
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 28
|
Chapter In summary, co-treatment with AlgL and dornase alfa seems to increase the efficacy of aminoglycosides in the presence of alginate. Treatment of Pa growing in biofilms can be improved by co-treatment with iron binding drugs, dispersion compounds or mannitol. dIsCussIon To our knowledge, this is the first systematic review on what happens to inhaled antibiot- ics after deposition in the airways of patients with CF. All results were primarily drawn from in vitro studies, from which we can conclude that the clinical efficacy of antibiotics is negatively affected by many factors after deposition in the airways ( Figure 2). Amino- glycosides, which were the most intensively studied relative to other inhaled antibiotics, seemed to be most adversely affected by these factors. Following dissolution, the drug needs to diffuse through the mucus layer to reach the site of the bacteria. The high concentration of macromolecules in mucus of patients with CF increases its viscosity, thereby impeding antibiotic diffusion. 87 Slow diffusion across CF mucus layers may play an important role in reduced pulmonary bioavailability as an- tibiotic molecules may be cleared by alveolar macrophages before reaching the bacteria. Therefore, coadministration of mannitol and dornase alfa may improve the diffusion of antibiotic molecules through mucus.90,Figure 2 – Pathway of the inhaled antibiotic after deposition on the mucus layer After depositing in the airways, the aerosol particle needs to dissolve in the airway surface layer or mucus layer. Next, the antibiotic needs to diffuse to the site where the bacteria are located. During the diffusion process through the mucus layer the aerosol particle can bind to molecules in the mucus. Also, the oxygen level, salt content and pH of the mucus are of influence on the antibiotic efficacy. Finally, the antibiotic has to overcome barriers generated by the microorganisms
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 29
|
Systematic review: fate of inhaled antibiotics During the diffusion process the inhaled antibiotic may bind to mucus, thereby limit - ing the amount of free drug available to be efficacious against bacteria. Aminoglycosides showed substantial binding to mucus from CF patients 94 while this was not observed in β-lactam antibiotics. This difference in binding can be attributed to the fact that aminoglycosides are positively charged, whereas β-lactam antibiotics are neutrally or negatively charged. As mucus macromolecules are negatively charged, they show a high affinity for positively charged antibiotics. This level of binding may be reduced by the coadministration of drugs such as cationic amphiphiles, which competitively bind to the macromolecules within the mucus. This resulted in saturation of the binding sites and a higher aminoglycoside bioavailability.93 Conflicting results on the coadministration of dornase alfa have been observed by different authors; one study found a decrease in binding,94 while another found an increase in binding.95 To the best of our knowledge, the effect of mucus on the efficacy of fluoroquinolones or macrolides has not been studied. Low oxygen levels and high salt concentrations within the mucus were shown to reduce the effectiveness of antibiotics. Specifically, aminoglycosides, β-lactam antibiot- ics, tetracyclines and chloramphenicol were rendered less efficacious against Pa under anaerobic conditions. Low oxygen levels within mucus may increase colistin activ- ity.80,107,109 With regard to salt concentrations, Pa was rendered less susceptible to killing by aminoglycosides in the presence of magnesium and calcium. 110 Additionally, nitrate decreased the efficacy of aminoglycosides and fluoroquinolones by half under anaerobic conditions.Ultimately, when antibiotic molecules make it to the vicinity of the bacteria, they still need to overcome multiple barriers generated by the microorganisms. The alginate layer surrounding Pa is an important contributing factor to its resistance of Pa against antibiotics and the patient's innate immune response. In general aminoglycosides bind to alginate while β-lactam antibiotics do not. Diffusion through alginate was impaired for all tested antibiotics, but most strongly for aminoglycosides. Coadministration of Algl showed promising results in vitro, with improved diffusion rates and enhanced bacteri - cidal activity.87,114 Another important barrier generated by Pa is biofilm formation, which drastically reduced effective killing. Treatment of Pa growing in biofilms can be further improved by co-treatment with iron binding glycoproteins, 116-118 mannitol or dispersion compounds.106,The limitations of this systematic review are the following; firstly, out of the 35 publica - tions selected for analysis, 9 publications were selected by screening the reference lists of the included articles. The search term "INHALED" was obligated in the title or abstract, while not all studies specified the route of administration. Secondly, a high level of inter-publication variability was observed for the following aspects of the studies; concentrations of antibiotic, types of alginate or exopolysaccha-
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 30
|
Systematic review: fate of inhaled antibiotics suPPlEmEnTAry TABlEs Table s1 – Critical appraisal table Criteria Additional information per criterion Study population CF patients (2), non-CF humans (1), animal study (0), in vitro/modelling study (0) Topics Direct answer or indirect answer to the research questions: direct (2), indirect (1), no answer (0) Study design experimental (2), observational (1), rest (0)* Clearly stated aim? Endpoints appropriate? Standardisation of outcome; clear definition of outcome measurement used Unbiased assessment of endpoint? Methods reproducible? Reporting bias Selective reporting Results objective? Conclusions justified? Additional criteria for comparative studies Selection bias Random sequence generation and allocation concealment: both (2), 1 out of 2 (1), none (0) Performance bias Blinding of participants and personnel Detection bias Blinding of outcome assessment Attrition bias Incomplete outcome data Adequate control group Contemporary groups Baseline equivalence If not: has there been a correction in the analysis? Follow-up appropriate? Loss to follow-up <5% If more loss to follow up: has selective loss been ruled out? Adequate statistics? Confounders Applicable? Sputum binding, concentration, abnormal CF sputum, resistance Total Score as 2 (reported adequately), 1 (reported inadequately/unclear), 0 (not reported), n/a (not appli- cable). * Experimental studies: RCT's, systematic reviews, meta-analysis. Observational studies: cohort studies, case-control studies, case series, case reports. Rest: animal studies, modelling studies, in vitro-studies
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 31
|
Chapter Table s2 – Critical appraisal completed Study population[1] 0 0 0 0 0 0 0 0 Topics[2] 1 1 1 1 2 1 1 1 Study design[3] 0 0 0 0 0 0 0 0 Clearly stated aim? 2 2 2 2 2 2 2 2 Endpoints appropriate?[4] 2 2 2 2 2 2 2 2 Unbiased assessment of endpoint? 2 2 2 2 2 2 0 0 Methods reproducible? 2 2 1 2 2 1 1 1 Reporting bias[5] 2 2 1 2 2 1 1 1 Results objective? 2 2 2 2 2 2 2 2 Conclusions justified? 2 2 2 2 2 2 2 2 Additional criteria for comparative studies Selection bias[6] Performance bias[7] Detection bias[8] Attrition bias[9] Adequate control group Contemporary groups Baseline equivalence[10] Follow-up appropriate? Loss to follow-up <5%[11] Adequate statistics? Confounders Applicable?[12] Total 15/20 15/20 13/20 15/20 16/20 13/20 11/20 11/20 15/Score as 2 (reported adequately), 1 (reported inadequately/unclear), 0 (not reported), n/a (not applicable). T = tobramycin, A = aztreonam [1] Study population: CF patients (2), non-CF humans (1), animal study (0), in vitro/modelling study (0); [2] Direct answer or indirect answer to the research questions: direct (2), indirect (1), no answer (0); [3] Study design: ex- perimental (2), observational (1), rest (0)*; [4] Standardisation of outcome; clear definition of outcome measure- ment used; [5] Selective reporting; [6] Random sequence generation and allocation concealment: both (2), 1 out of 2 (1), none (0); [7] Blinding of participants and personnel; [8] Blinding of outcome assessment; [9] Incomplete outcome data; [10] If not: has there been a correction in the analysis?; [11] If more loss to follow up: has selective loss been ruled out?; [12] Sputum binding, concentration, abnormal CF sputum, resistance. * Experimental studies: RCT's, systematic reviews, meta-analysis. Observational studies: cohort studies, case- control studies, case series, case reports. Rest: animal studies, modelling studies, in vitro-studies.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 32
|
Systematic review: fate of inhaled antibiotics Continuation Study population[1] 0 0 0 0 0 0 0 0 Topics[2] 1 1 1 1 2 2 1 1 Study design[3] 0 0 0 0 0 0 0 0 Clearly stated aim? 2 2 2 2 2 2 2 2 Endpoints appropriate?[4] 2 2 2 2 2 2 2 2 Unbiased assessment of endpoint? 2 2 0 2 2 2 2 2 Methods reproducible? 2 2 1 2 2 2 2 2 Reporting bias[5] 2 2 1 2 2 2 2 1 Results objective? 2 2 2 2 2 2 2 1 Conclusions justified? 2 2 2 2 2 2 2 1 Additional criteria for comparative studies Selection bias[6] Performance bias[7] Detection bias[8] Attrition bias[9] Adequate control group Contemporary groups Baseline equivalence[10] Follow-up appropriate? Loss to follow-up <5%[11] Adequate statistics? Confounders Applicable?[12] Total 15/20 15/20 11/20 15/20 16/20 16/20 15/20 12/20 15/Score as 2 (reported adequately), 1 (reported inadequately/unclear), 0 (not reported), n/a (not applicable). T = tobramycin, A = aztreonam [1] Study population: CF patients (2), non-CF humans (1), animal study (0), in vitro/modelling study (0); [2] Direct answer or indirect answer to the research questions: direct (2), indirect (1), no answer (0); [3] Study design: ex- perimental (2), observational (1), rest (0)*; [4] Standardisation of outcome; clear definition of outcome measure- ment used; [5] Selective reporting; [6] Random sequence generation and allocation concealment: both (2), 1 out of 2 (1), none (0); [7] Blinding of participants and personnel; [8] Blinding of outcome assessment; [9] Incomplete outcome data; [10] If not: has there been a correction in the analysis?; [11] If more loss to follow up: has selective loss been ruled out?; [12] Sputum binding, concentration, abnormal CF sputum, resistance. * Experimental studies: RCT's, systematic reviews, meta-analysis. Observational studies: cohort studies, case- control studies, case series, case reports. Rest: animal studies, modelling studies, in vitro-studies.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 33
|
Chapter Continuation Study population[1] 0 0 0 0 0 0 0 0 Topics[2] 1 1 1 1 1 2 2 2 Study design[3] 0 0 0 0 0 0 0 0 Clearly stated aim? 2 2 1 2 1 2 2 2 Endpoints appropriate?[4] 2 2 2 2 1 2 2 2 Unbiased assessment of endpoint? 2 2 2 2 2 2 2 2 Methods reproducible? 2 2 2 2 2 2 2 2 Reporting bias[5] 2 2 2 2 2 2 2 1 Results objective? 2 2 2 2 2 2 2 2 Conclusions justified? 2 2 2 2 2 2 2 2 Additional criteria for comparative studies Selection bias[6] Performance bias[7] Detection bias[8] Attrition bias[9] Adequate control group Contemporary groups Baseline equivalence[10] Follow-up appropriate? Loss to follow-up <5%[11] Adequate statistics? Confounders Applicable?[12] Total 15/20 15/20 14/20 15/20 13/20 16/20 16/20 15/20 16/Score as 2 (reported adequately), 1 (reported inadequately/unclear), 0 (not reported), n/a (not applicable). T = tobramycin, A = aztreonam [1] Study population: CF patients (2), non-CF humans (1), animal study (0), in vitro/modelling study (0); [2] Direct answer or indirect answer to the research questions: direct (2), indirect (1), no answer (0); [3] Study design: ex- perimental (2), observational (1), rest (0)*; [4] Standardisation of outcome; clear definition of outcome measure- ment used; [5] Selective reporting; [6] Random sequence generation and allocation concealment: both (2), 1 out of 2 (1), none (0); [7] Blinding of participants and personnel; [8] Blinding of outcome assessment; [9] Incomplete outcome data; [10] If not: has there been a correction in the analysis?; [11] If more loss to follow up: has selective loss been ruled out?; [12] Sputum binding, concentration, abnormal CF sputum, resistance. * Experimental studies: RCT's, systematic reviews, meta-analysis. Observational studies: cohort studies, case- control studies, case series, case reports. Rest: animal studies, modelling studies, in vitro-studies.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 34
|
Systematic review: fate of inhaled antibiotics Continuation Study population[1] 0 0 0 0 0 0 0 Topics[2] 2 1 2 2 2 1 2 Study design[3] 0 0 0 0 0 0 0 Clearly stated aim? 1 2 2 2 2 2 2 Endpoints appropriate?[4] 2 2 2 2 2 2 2 Unbiased assessment of endpoint? 2 2 2 2 2 2 2 Methods reproducible? 2 2 1 2 2 2 1 Reporting bias[5] 2 2 2 2 2 2 2 Results objective? 2 2 2 2 2 2 2 Conclusions justified? 2 1 1 2 2 2 2 Additional criteria for comparative studies Selection bias[6] Performance bias[7] Detection bias[8] Attrition bias[9] Adequate control group Contemporary groups Baseline equivalence[10] Follow-up appropriate? n.a. Loss to follow-up <5%[11] n.a. Adequate statistics? Confounders Applicable?[12] Sputum binding Total 15/20 14/20 14/20 16/20 20/38 15/20 15/20 15/Score as 2 (reported adequately), 1 (reported inadequately/unclear), 0 (not reported), n/a (not applicable). T = tobramycin, A = aztreonam [1] Study population: CF patients (2), non-CF humans (1), animal study (0), in vitro/modelling study (0); [2] Direct answer or indirect answer to the research questions: direct (2), indirect (1), no answer (0); [3] Study design: ex- perimental (2), observational (1), rest (0)*; [4] Standardisation of outcome; clear definition of outcome measure- ment used; [5] Selective reporting; [6] Random sequence generation and allocation concealment: both (2), 1 out of 2 (1), none (0); [7] Blinding of participants and personnel; [8] Blinding of outcome assessment; [9] Incomplete outcome data; [10] If not: has there been a correction in the analysis?; [11] If more loss to follow up: has selective loss been ruled out?; [12] Sputum binding, concentration, abnormal CF sputum, resistance. * Experimental studies: RCT's, systematic reviews, meta-analysis. Observational studies: cohort studies, case- control studies, case series, case reports. Rest: animal studies, modelling studies, in vitro-studies.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 35
|
Chapter Continuation 4 (Excluded articles) Study population[1] 0 0 1 1 0 1 1 2 0 Topics[2] 1 1 2 0 1 1 1 0 2 Study design[3] 0 0 1 1 1 0 1 0 0 Clearly stated aim? 2 1 2 2 2 1 1 2 2 Endpoints appropriate?[4] 2 0 1 1 2 1 1 2 2 Unbiased assessment of endpoint? 1 1 0 1 1 1 1 2 2 Methods reproducible? 0 1 1 0 0 0 0 2 2 Reporting bias[5] 1 0 1 1 1 1 1 2 2 Results objective? 1 1 1 1 1 1 1 2 2 Conclusions justified? 1 1 2 2 1 2 2 2 2 Additional criteria for comparative studies Selection bias[6] Performance bias[7] Detection bias[8] Attrition bias[9] Adequate control group Contemporary groups Baseline equivalence[10] Follow-up appropriate? n.a. Loss to follow-up <5%[11] n.a. Adequate statistics? Confounders Applicable?[12] Review Solely Pa Total 9/20 6/20 12/20 10/20 10/20 9/20 10/20 26/38 16/20 1/Score as 2 (reported adequately), 1 (reported inadequately/unclear), 0 (not reported), n/a (not applicable). T = tobramycin, A = aztreonam [1] Study population: CF patients (2), non-CF humans (1), animal study (0), in vitro/modelling study (0); [2] Direct answer or indirect answer to the research questions: direct (2), indirect (1), no answer (0); [3] Study design: ex- perimental (2), observational (1), rest (0)*; [4] Standardisation of outcome; clear definition of outcome measure- ment used; [5] Selective reporting; [6] Random sequence generation and allocation concealment: both (2), 1 out of 2 (1), none (0); [7] Blinding of participants and personnel; [8] Blinding of outcome assessment; [9] Incomplete outcome data; [10] If not: has there been a correction in the analysis?; [11] If more loss to follow up: has selective loss been ruled out?; [12] Sputum binding, concentration, abnormal CF sputum, resistance. * Experimental studies: RCT's, systematic reviews, meta-analysis. Observational studies: cohort studies, case- control studies, case series, case reports. Rest: animal studies, modelling studies, in vitro-studies.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 36
|
Systematic review: fate of inhaled antibiotics Table s3 – Main results according to step in antibiotic pathway Step in antibiotic pathway number of studies Main results 1. dissolution in mucus layer 2 studies describing the same in vitro study • Tobramycin dissolves easier in mucus of secretors → more rapid effect (2 studies) 2. diffusion through mucus layer 4 in vitro studies • Aminoglycosides (1 study), β-lactam (1 study), Fluoroquinolones (2 studies): CF mucus reduces antibiotic diffusion 3. Binding to molecules in mucus 8 in vitro studies • Aminoglycosides (7 studies): o Strong binding to mucin (2 studies), DNA (4 studies), nucleic acids (2 studies) and bacterial endotoxins (LPS, LTS) (1 study). Strong binding mucus in general (3 studies). o Binding tobramycin 15-95% (2 studies), amikacin 1-60% (1 study), gentamicin 52% (1 study) o Liposomal form: binding reduced 4 to 100-fold (1 study) o Dornase alfa treatment: ▪ Increased binding (1 study) ▪ Decreased binding (1 study) • β-lactam AB: negligible binding to CF mucus (2 studies) • Polymyxin B (2 studies): o Strong binding to mucus in general (1 study). Binding to bacterial endotoxins, but no binding to DNA/F-actin (1 study) o Liposomal form: binding reduced up to 100-fold (1 study) 4. Influence of oxygen level in mucus 4 in vitro studies • Aminoglycosides (3 studies), β-lactam AB (3 studies), Macrolides (1 study), Cotrimoxazol (1 study), Tetracyclines (1 study), Chloramphenicol (1 study): Reduced efficacy in anaerobic conditions. • Fluoroquinolones: o Reduced efficacy in anaerobic conditions (2 studies) o Remain bactericidal in anaerobic conditions (1 study) • Colistin: o Reduced efficacy in anaerobic conditions (1 study) o Increased efficacy in anaerobic conditions (1 study) 5. Influence of salt content in mucus 2 in vitro studies • Aminoglycosides: efficacy reduced by salts (2 studies), high ionic strength (polyanions) (1 study) • Fluoroquinolones: susceptibility of organisms to AB decreased by nitrate (1 study) • β-lactam AB (1 study): o Carbenicillin: nitrate increased efficacy if aerobic, but effect abolished if anaerobic o Ceftazidime: nitrate no effect • Chloramphenicol and tetracycline: nitrate no effect (1 study)
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 37
|
Chapter Table s3 – Main results according to step in antibiotic pathway (continued) Step in antibiotic pathway number of studies Main results 6. diffusion through Pa alginate layer 10 in vitro studies • Aminoglycosides: o Inhibited in diffusion by alginate, stronger inhibition than β-lactam AB (3 studies) o 12-fold reduction in aminoglycoside concentration in 1.5% w/v Pa alginate (1 study) o Decrease in aminoglycoside binding to alginate in presence of salt (1 study) • β-lactam AB: inhibited in diffusion by alginate (1 study) • Polymyxin B: inhibited in diffusion by alginate (1 study) 7. liposomal formulations 3 in vitro studies, 2 animal studies • Aminoglycosides: o Liposomal form more efficacious than free form (4 studies) o Liposomal form less efficacious than free form (1 study) • Polymyxin B: liposomal form more efficacious than free form (1 study) 8. Co-treatment with other medications 15 in vitro studies • Dornase alfa: o Enhanced bactericidal activity of aminoglycosides (2 studies) o Reduced bactericidal activity of aminoglycosides (1 study) o Enhanced bactericidal activity of fluoroquinolones (1 study) • Mannitol: enhanced bactericidal activity of aminoglycosides (1 study) and fluoroquinolones (2 studies) • Alginate lyase: improved diffusion rates and enhanced bactericidal activity of aminoglycosides (2 studies), combination with dornase alfa most effective • N-acetylcysteine: no effect on bactericidal activity of aminoglycosides • NaCl: o Synergistic effect on colistin (1 study) o Enhanced bactericidal activity of tobramycin (1 study) o Antagonistic effect on tobramycin (1 study) • Glucose: enhanced bactericidal activity of aminoglycosides (1 study) • ALX-109&ALX009: enhanced bactericidal activity of aminoglycosides (1 article, 1 abstract) and β-lactam AB (1 article, 1 abstract) • L-leucine: no influence on aminoglycoside permeability through mucus (1 study) • Cationic amphiphiles: reduced binding between aminoglycosides and DNA (1 study) • Dispersion compounds: o Synergistic action of citrate with amikacin, colistin or erythromycin and succinic acid with colistin (1 study) o Reduced bactericidal activity of tobramycin and polymyxin B (1 study) AB = antibiotic; CF = cystic fibrosis; LPS = lipopolysaccharides; LTA = lipoteichoic acid.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 38
|
Chapter ABsTrACT Background: Pseudomonas aeruginosa (Pa) infection is an important contributor to the progression of cystic fibrosis (CF) lung disease. The cornerstone treatment for Pa infection is the use of inhaled antibiotics. However, there is substantial lung disease heterogeneity within and between patients that likely impacts deposition patterns of inhaled antibiot- ics. Therefore, this may result in airways below the minimal inhibitory concentration of the inhaled agent. Very little is known about antibiotic concentrations in small airways, in particular the effect of structural lung abnormalities. We therefore aimed to develop a patient-specific airway model to predict concentrations of inhaled antibiotics and to study the impact of structural lung changes and breathing profile on local concentrations in airways of patients with CF. methods: In- and expiratory CT-scans of children with CF (5-17 years) were scored (CF-CT score), segmented and reconstructed into 3D airway models. Computational fluid dy - namic (CFD) simulations were performed on 40 airway models to predict local Aztreonam lysine for inhalation (AZLI) concentrations. Patient-specific lobar flow distribution and nebulization of 75 mg AZLI through a digital Pari eFlow model with mass median aero- dynamic diameter range were used at the inlet of the airway model. AZLI concentrations for central and small airways were computed for different breathing patterns and airway surface liquid thicknesses. results: In most simulated conditions, concentrations in both central and small airways were well above the minimal inhibitory concentration. However, small airways in more diseased lobes were likely to receive suboptimal AZLI. Structural lung disease and in- creased tidal volumes, respiratory rates and larger particle sizes greatly reduced small airway concentrations. Conclusions: CFD modeling showed that concentrations of inhaled antibiotic delivered to the small airways are highly patient specific and vary throughout the bronchial tree. These results suggest that anti- Pa treatment of especially the small airways can be improved.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 39
|
Modeling of antibiotic concentrations in CF patients InTroduCTIon Cystic fibrosis (CF) is a severe hereditary and life-threatening disease in the Caucasian population. Most morbidity and mortality (>90% of deaths) is caused by progressive lung disease. 10 Important components of the pathophysiology of CF lung disease are bronchiectasis, small airways disease 5-7,119 and chronic infection with Pseudomonas aeruginosa (Pa) as the main pathogen.Inhaled antibiotics play a central role in eradication and chronic suppressive therapy of Pa infections. Unfortunately, despite these interventions, lung disease in CF eventu- ally progresses to end-stage, with substantial small airways disease in most patients. 9 Significant mucus accumulation and wall thickening in the small airways has been found in explant CF lungs with end-stage lung disease 120,121 and this has been associated with the presence of Pa.122 Hence, more effective anti-Pa therapies, especially those targeted at the small airways, may offer an opportunity to improve patient outcomes. The generally held view that inhaled antibiotics result in high concentrations within the airways is largely based on high drug concentrations found in sputum. 123,124 How- ever, it is unlikely that sputum concentrations are representative for the small airway concentrations. The drug reaching the small airways is distributed over a much larger surface area, namely a 30-190-fold greater area compared to central airways.125 In addi- tion, mucociliary transport clears sputum from the small airways via the central airways, taking up additional drug during transit, before expectoration. Thus, the final sputum concentration is likely to overestimate small airway concentration. Very little is known about antibiotic concentrations in the small airways, due to the difficulty of in vivo measurement. The progression of small airways disease despite anti-Pa treatment suggests that small airway deposition of inhaled antibiotics may be insufficient. To optimize Pa eradication and chronic suppression with inhaled antibiotics, it is important to obtain local concentrations equal to or above the minimal inhibitory concentration (MIC). Concentrations below the MIC lead to the development of Pa strains with high mutation rates, and hence resistant subpopulations of Pa which cannot be eradicated.Extensive research has been done to understand aerosol deposition mechanisms. It has been well established that aerosol deposition is strongly dependent on particle size,127 airflow, inhalation technique, lung structural changes and airway obstruction by mucus.128 CF-patients with more severe lung disease have more central airway deposi- tion compared to healthy individuals.128 This suggests that dose adjustments and particle size optimization, or inhalation technique, could improve aerosol delivery to the site of infection. However, to maximize drug delivery to the small airways, the impact of age, structural changes and inspiratory flow profile on antibiotic concentrations in different compartments of the bronchial tree needs to be better understood.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 40
|
Chapter Unfortunately, it is difficult to investigate the simultaneous influence of the above- mentioned factors on deposition in vivo. An in silico, patient-specific model based on computational fluid dynamics (CFD) has been developed to assess the behavior of inha- lation medication in airways. 129 This technique has been validated using Single Photon Emission Computed Tomography.130 To date, this technique has been used to study lung drug deposition in asthma,63 to assess airflow distribution in both asthma and chronic ob- structive pulmonary disease,129,130 and the bronchodilating effects of β2-agonists.131-In CF, CFD can allow us to study the relation between airway morphology and local concentrations of inhaled antibiotics. Additionally, by repeating simulations with various model parameters, CFD can provide more information on how to optimize small airway aerosol deposition in CF-patients with structural lung changes. This is the first study using patient-specific airway models with varying disease sever- ity and CFD to estimate aerosol concentrations in both the central and small airways of patients with CF. We aimed to study the relation between structural lung disease and deposition of an inhaled antibiotic used for suppressive treatment of chronic Pa infec- tions, Aztreonam lysine for inhalation (AZLI; Gilead Pharmaceuticals, Foster City, USA). AZLI is a monobactam antibiotic, delivered by the e-Flow electronic nebulizer. 124 We hypothesized that: a) there is great variation in AZLI concentrations between patients, due to differences in airway geometry and lung disease severity, b) AZLI concentrations in the small airways would be below the MIC for Pa in patients with more severe lung disease, and c) AZLI concentrations in the small airways could be improved by increasing the dose of AZLI or by modifying the inhalation technique. mATErIAls And mEThods Study population We included all spirometer controlled volumetric in- and expiratory high resolution CT- scans, with a slice thickness of 1 mm or less, performed as part of the routine annual CF check-up in the CF-centre of Erasmus MC-Sophia Children's Hospital (Rotterdam, the Netherlands) between 2008 and 2012 (aged 5-17 years). Patients were diagnosed with CF by a positive sweat test and/or genotyping for known CF mutations. Demographic data and pulmonary function tests were collected prior to the CT-scan. Pulmonary function test results were expressed as percentages of predictive values, according to Stanojevic for the forced vital capacity (FVC) and forced expiratory volume in 1 second (FEV 1), and Zapletal for the forced expiratory flow at 75% (FEF75).134,135 Written informed consent for the use of de-identified data was obtained from the parent/guardian and subjects ≥ 12
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 41
|
Modeling of antibiotic concentrations in CF patients years. This retrospective study was approved by the Institutional Review Board of the Erasmus Medical Center in Rotterdam, the Netherlands (MEC-2013-078). Chest Computed Tomography (CT) Forty spirometer controlled CT-scans of consecutive CF-patients, acquired as part of routine clinical care, were included. To quantify chest CT abnormalities, we used the vali- dated CF-CT scoring system.136 The lobar specific CF-CT score per component was used and expressed as a percentage of the maximum possible score per lobe. The component scores for bronchiectasis, airway wall thickening and air trapping were used for analysis. Detailed descriptions of CT scanning protocol and CT evaluation are available in the sup- porting information of this paper (Text S1). reconstruction of three-dimensional airway models Based on the inspiratory scan, a semi-automatic algorithm was used to reconstruct a patient-specific three-dimensional (3D) model of the intra-thoracic region. This intra- thoracic region was defined arbitrarily as the lower airway. Automatic airway segmenta - tion was performed up to the point where no distinction could be made between the intra-luminal and alveolar air. Following automated segmentation of the bronchial tree, the airways were manually checked. Missing branches were added to the bronchial tree and incorrect branches were deleted when necessary; 3.39±2.51% of the branches needed to be manually altered. The respiratory tract was reconstructed down to the level of airways with a diameter of 1–2mm. The segmented airway tree was converted into a 3D model that was smoothed using a volume compensation algorithm. The smoothed model was trimmed perpendicular to the airway centreline at the trachea (using the middle point of the superior side of the sternum as a landmark) and at each terminal bronchus. Remaining artefacts due to noise in the CTs were then manually removed from the model. For the upper (extra-thoracic) airways, a generic average adult upper airway model was selected and scaled down in such a way that both the anteroposterior and lateral dimension of the scaled model's trachea, at the location of the sternum, matched the average anteroposterior (1.25cm) and lateral (1.19cm) dimension for the 40 patients. The upper airway model was connected with a reverse engineered mouthpiece of the Pari eFlow. Reverse engineering was done based on a CT-scan of the mouthpiece taken on a GE LightSpeed VCT (80kV, 18.25mAs, 0.311mm slice increment, 0.188mm pixel size, STANDARD reconstruction algorithm). The mouthpiece/upper airway model was trimmed perpendicular to the centreline of the trachea (again using the middle point of the superior side of the sternum as a landmark). This ensured correct positioning of the upper airway model with respect to the patient-specific airway models. Models were then coupled using the freeform hole filling algorithm of 3-Matic.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 42
|
Chapter For each of the 40 CT-scan sets, the patient-specific lower airway model was con - nected to the selected nebulizer mouthpiece/upper airway model, maximizing the contribution of patient-specific information. All segmentation and 3D model operations were performed in commercially available validated software packages (Mimics 15.0 and 3-Matic 7.0, Materialise N.V., Belgium, Food and Drug Administration, K073468; Confor- mité Européenne certificate, BE 05/1191.CE.01). meshing The triangulated, mouthpiece/upper/lower airway surface models had a maximum tri- angle edge length of 0.5 mm and a minimum triangle aspect ratio of 0.4. These models were converted to tetrahedral 3D volume meshes using TGrid 14.0 (Ansys Inc, Canons- burg, PA). A boundary layer with a growth rate of 1.4 was included in the models. Maximal tetrahedral volume was set to 2 mm 3 and maximal equilateral volume-based skewness to 0.9. Grid convergence demonstrated that a mesh size of 2.9±0.7M [1.9–4.6] cells is appropriate for the study, depending on the size of the patient-specific lower airway model. Meshing was done on 1 CPU and meshing time was below 200s. reconstruction of three-dimensional lung lobes From both the inspiratory and expiratory CT-scans, the patient-specific lung lobes were extracted using a semi-automated tool that identifies the fissures separating the lung lobes. The internal lobar flow distribution was calculated based on the lobar volume change from expiration to inspiration. Lung lobe identification has been performed in a commercially available validated software package (Mimics 15.0, Materialise N.V., Belgium, Food and Drug Administration, K073468; Conformité Européenne certificate, BE 05/1191.CE.01). Inlet of the airway model Breathing profile. The median age of the patient population (11 years) was used to gen- erate a generic breathing profile based on the following parameters: the median weight of 11 year old Dutch children is 38 kg (boy: 37 kg, girl 38.5 kg) 137; tidal volume of 10 ml/kg (380 ml); respiration rate (18 breaths per minute). 138 The resulting profile had an inspiration/expiration ratio of 1:2 and a sinusoidal shape, see S1 Fig. To be able to examine the flow dependency of the simulated results, two additional breathing profiles were generated: (1) a high breathing profile, consisting of a higher tidal volume of 14 ml/kg (532 ml) and the respiratory rate of the youngest age (5 years: 22 breaths per minute), and (2) a low breathing profile, consisting of a lower tidal volume of 6 ml/kg (228 ml) and the respiratory rate of oldest age (17 years: 14 breaths per minute). These additional profiles can also be found in S1 Fig.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 43
|
Modeling of antibiotic concentrations in CF patients Aerosol characteristics. Eleven different trials (Anderson Cascade impactor n=6, next generation impactor n=2, and laser diffraction n=3) studied the diameter distribution of AZLI nebulized via the Pari eFlow (Gilead data on file). The extremes and the median of these unpublished trials were selected for use in the CFD simulations: smallest diameter (2.81±1.47 μm), median diameter (3.18±1.63μm) and largest diameter (4.35±2.05 μm). Furthermore, an in vivo characterization of the eFlow showed that 35% of the nominal fill volume is either trapped in the mouthpiece or exhaled. Flow simulation. Computational fluid dynamics (CFD) flow simulations were per - formed in Fluent 14.0 (Ansys Inc, Canonsburg, PA). Drug release in the simulated nebulizer was continuous. Therefore, particles were injected during the whole breathing cycle. All simulations were transient using a second order time-stepping algorithm and a time-step of 0.005s. Turbulence was evaluated through large eddy simulations with a turbulent kinetic subgrid model. Aerosol transport was modeled by an implicit Runge-Kutta La- grangian discrete particle model, with a one-way coupling of the forces from the flow to the particle and taken into account the Saffman lift forces. Transient particle tracking was used and the particle time-step was equal to the flow time-step. Every time-step, 15862 particles were injected. This number is based on particle convergence studies. Particles were considered deposited the moment they hit the airway wall. The nominal dose of 75 mg AZLI was corrected for the 35% combined inhaler loss and exhaled fraction (Gilead data on file). Due to the incorporation of the exhaled fraction, only the inhalation was modeled. The boundary condition at the inhaler mouthpiece was represented by the inhalation part of the mean breathing profile in S1 Fig. The down - stream boundary conditions at the terminal bronchi were set such that the percentage of flow exiting the model towards a lobe did match with the internal lobar flow distribution obtained from the expiratory and inspiratory CT data. To investigate the influence of the inhalation manoeuvre on local concentrations, ad- ditional simulations were performed in a subset of the population (2 tallest, 2 smallest, 2 median sized patients). These additional CFD simulations were performed with the altered breathing profiles described in Section: 'Breathing profiles'. Calculation of regional AZlI concentrations Aerosol deposition analyses. To be able to perform regional analyses, the respiratory tract was subdivided into multiple regions. For the airways with a diameter >1–2 mm these regions were obtained from the mouthpiece/upper/lower airway model, see Fig. 1. In this figure the upper airway is divided in two parts: the oral cavity and the pharynx; and the lower airways are divided into central part and distal parts representing the lung segments. Conducting airways with a diameter <1–2 mm could not be distinguished from the CT images and have been added to the patient-specific model by using Phalen's
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 44
|
Chapter description of the airway tree in infants, children and adolescents. 125 For every simula- tion, a Phalen model was constructed based on the height of the specific patient. Regional AZLI deposition was evaluated for both the particles depositing inside the model, in every separate zone indicated in Fig. 1; as well as for the particles exiting the model at the terminal bronchi, in the small airways represented by the Phalen model on a lobar basis. Once the aerosol entered the Phalen model of a certain lobe, it was assumed that it was distributed homogeneously. Airway surface liquid. To compute AZLI concentrations in the airway surface liquid (ASL) throughout the bronchial tree we used a range of thicknesses based on studies in CF. Three different ASL scenarios were considered: thick ASL (7 µm),139 thin ASL (3 μm)140 and the mean ASL (5 μm). AZlI concentrations. For each reconstructed airway and for each lung lobe, the area was calculated and the CFD simulations provided data on the drug deposition in that region. The regional AZLI concentration was computed as follows: the mass of the depos- ited drug in an airway was divided by the thickness of the lining fluid multiplied by the surface area of that airway. Figure 1 – Coupled mouthpiece/upper/lower airway model Coupled mouthpiece/upper/lower airway model subdivided in multiple regions. Airways are seg- mented up to the 5th-9th generation.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 45
|
Chapter rEsulTs Study population Forty inspiratory and expiratory chest CT-scans were selected from 31 patients. Baseline characteristics are shown in Table 1. Thirty-nine (98%) CT-scans were spirometer con- trolled; the remaining scan was performed with technician guidance. There were no significant differences between the sexes for demographics, pulmonary function tests and CF-CT subscores, therefore the dataset did not have to be split in sex groups. ICCs for within-observer agreement ranged from 0.85 (air trapping) to 0.93 (bronchiec- tasis), whereas between-observer agreement ranged from 0.67 (airway wall thickening) to 0.77 (air trapping). Table 1. Baseline characteristics. Value N - Nr of patients with 1 CT - Nr of patients with 2 CTs Male 11 35% Age 11.0 5.8-17.Bronchiectasis score (% of max CF-CT score) 2.8 0.0-16.Airway wall thickening score (% of max CF-CT score) 3.7 0.0-18.Air trapping score (% of max CF-CT score) 22.2 11.1-85.FEV1 %pred 94.2 70.8-115.FVC %pred 104.3 78.7-127.Data are presented as nr. (%) or median (range), unless otherwise indicated. Correlations with age There was a moderate positive correlation between bronchiectasis and age (r s=0.481, p=0.005). Correlations between age and airway wall thickness (r s=0.302, p=0.087), air trapping (r s=0.096, p=0.554) and pulmonary function tests (FVC% pred: r s=0.053, p=0.885; FEV1% pred: rs=-0.024, p=0.885) were not significant. Deposition analyses The software tool used to identify the boundaries between the lobes in the lungs, i.e. the pulmonary fissures, could not identify the fissure between the right upper and middle lung lobes in 18 patients and between the left upper and lower lung lobes in 1 patient. These lung lobes were excluded for analysis of AZLI deposition.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 46
|
Modeling of antibiotic concentrations in CF patients Significant differences between the lobes were found for all tested CF-CT subscores (bronchiectasis: x2 =21.70, p<0.001; airway wall thickness: x 2 =25.22, p<0.001; and air trapping: x2 =20.15, p<0.001). It was found that CF-CT subscores were generally higher in the right upper lobe than in the other lobes (Fig. 2). There were differences in AZLI deposition between the different lobes ( Fig. 3). The highest AZLI concentrations were found in the lower lobes. For the lower lobes, AZLI con- centrations were always above 10xMIC90 independent of the scenario tested. An inverse correlation between AZLI concentration in a lobe and the CF-CT scores was observed, indicating that more diseased lobes received less drug (Table 2). For example, when as- suming small diameters and thin lining fluids, a reduction in AZLI concentration of 439 μg/ml was observed for every 1% of point increase in bronchiectasis score. AZLI concentrations were calculated with 3 different ASL thicknesses. Because of the formula used for calculations, the ASL thickness was of direct influence on the concentrations: the thicker the ASL the lower the AZLI concentration (S2 Fig.). Therefore, when expressing this regional AZLI concentration relative to 10xMIC 90, the thicker the ASL the larger the area of small airways with AZLI concentrations below 10xMIC 90. Figure 2 – Comparison of CF-CT subscores per lobe Comparison of CF-CT subscores per lobe, presented as % of max CF-CT score. Data are presented as median (range), unless otherwise indicated. White bars represent bronchiectasis score, light grey bars represent airway wall thickening score and dark grey bars represent air trapping score. RUL = right upper lobe (n=22), RML = right middle lobe (n=22), RLL = right lower lobe (n=40), LUL = left upper lobe (n=39), LLL = left lower lobe (n=39).
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 47
|
Chapter Figure 3 – differences between lobes in AZlI concentrations Differences between lobes in AZLI concentrations for the scenario of thick airway surface liquid with largest aerosol diameter. Data are presented as median (range), unless otherwise indicated. Significant differences in AZLI concentrations were found between all lobes, except for one pairwise comparison (see Table S1). RUL = right upper lobe, RML = right middle lobe. RLL = right lower lobe, LUL = left upper lobe, LLL = left lower lobe. Figure 4 – Percentage area of small airways with AZlI <10xmICPercentage area of small airways with AZLI concentrations <10xMIC90. Data are presented as median (range) for the different scenarios. White bars represent the smallest aerosol diameter (2.9 µm), light grey bars represent the median aerosol diameter (3.18 μm) and dark grey bars represent the largest aerosol diameter (4.35 μm). ASL = airway surface liquid.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 48
|
Chapter The combination of thin ASL and smallest aerosol diameter resulted in AZLI concentra- tions above 10xMIC90 for both large and small airways. For the combination of thick ASL and largest aerosol diameter, 22% (0-49.79%) of the total area of small airways received AZLI concentrations below 10xMIC 90. The lowest AZLI value observed in the small air- ways for the tested population was 468.14 μg/ml or 3.66xMIC 90. Fig. 4 summarizes the percentage of area of small airways that receive a concentration below 10xMIC90 for the different modeling conditions. In Fig. 5, the relative AZLI concentrations in 2 patients are shown for 3 different scenarios. In the central and distal airways, AZLI concentrations 10- 100x above the threshold of 1280 μg/ml were observed. In the small airways (visualized per lobe in the images), lower concentrations were seen. Decreasing the tidal volume and respiratory rates decreased the deposition in the extra-thoracic region (mouth and upper airway), subsequently resulting in significantly less areas with a concentration below 10xMIC90 in the lungs (Fig. 6). Figure 5 – relative AZlI concentrations in central and small airways of 2 patients for 3 different scenarios Simulations of AZLI deposition in 2 patients, representing 3 scenarios of varied airway surface liquid thickness (ASL) and aerosol diameter. Severity of CF lung disease was determined by the CF-CT score (% of total CF-CT score). Scenario a = thin ASL with smallest aerosol diameter; scenario b = median ASL with median aerosol diameter; scenario c = thick ASL with largest aerosol diameter. Part 1a, 1b and 1c: Patient 1, mild CF lung disease: bronchiectasis 0.0%, airway wall thickening 0.0% and air trapping 11.1%. Patient 1 received concentrations > 10xMIC90 in the central and small airways inde- pendent of ASL thickness and aerosol diameter (Part 1a, 1b, 1c). Part 2a, 2c and 2c: Patient 2, more severe lung disease: bronchiectasis 12.5%, airway wall thickening 11.1% and air trapping 38.9%. Patient 2 received concentrations > 10xMIC 90 in the central and small airways in scenario a and b (Part 2a and 2b), but AZLI concentrations < 10xMIC90 in the small airways in scenario c (right upper and middle lobes) (Part 2c).
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 49
|
Modeling of antibiotic concentrations in CF patients dIsCussIon To our knowledge this is the first study that used CFD to estimate patient-specific inhaled antibiotic concentrations throughout the bronchial tree in CF. The most important finding was that the airway concentrations were highly dependent on patient-related factors. Another important finding of this study was that effective AZLI concentrations above the 10xMIC 90 threshold for Pa were observed throughout the lung in most simulated conditions. However, variables such as particle diameter and ASL thickness had a sig- nificant impact on the results. Under certain values of these variables, it was shown that the concentration would drop below 10xMIC 90 in 22% of the small airways area. The most critical scenario was the combination of a thick ASL and the largest aerosol diameter. However, the lowest observed AZLI concentration in the small airways of the studied population was still above 3xMIC90 for Pa. For this particular patient, an increase Figure 6 – Influence of inhalation technique on AZlI concentrations Influence of inhalation technique on AZLI concentrations presented as percentage area of small airways with AZLI <10xMIC90. Low breathing profile: tidal volume of 6 ml/kg (228 ml) and respiration rate of 14 breaths/min. Average breathing profile: tidal volume of 10 ml/kg (380 ml) and respiration rate of 18 breaths/min. High breathing profile: tidal volume of 14 ml/kg (532 ml) and respiration rate of 22 breaths/min. Data are presented as median (range) for the different scenarios. Light grey bars represent the scenario of median ASL (5 μm) with largest aerosol diameter (4.35 μm). The darker grey bars represent the scenario of thick ASL (7 μm) with median aerosol diameter (3.18 μm) and the darkest grey bars represent the scenario of thick ASL (7 μm) with largest aerosol diameter (4.35 μm). The scenarios of thin ASL with all diameters, median ASL with smallest and median diameter and thick ASL with smallest diameter are not represented as all breathing profiles resulted in AZLI concentrations above 10xMIC90. ASL = airway surface liquid.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 50
|
Chapter of 2.7 times the standard nebulized AZLI dose would have resulted in sufficiently high concentrations in the small airways. In the 'best-case' scenario, both large and small airways received AZLI concentrations above 10xMIC90. The observation that regional low concentrations can exist is of great importance, since suboptimal concentrations could result in insufficient killing and are associated with increases in mutation frequencies.144 These hypermutator strains are resistant against antimicrobials used in CF and hence, rather than elimination of the pathogen, treatment will result in even further selection of these resistant subpopulations.As in previous studies we observed that the upper lobes were more severely affected by structural disease relative to the other lobes.119 The reason for this distribution is still largely unknown, however our findings suggest that uneven distribution of inhaled drugs could contribute to this inequality. We observed that even in patients with relatively little structural damage, the upper lobes received lower AZLI concentrations than the lower lobes. We found an inverse correlation between lobar CF-CT score and AZLI concentration. The population included in this study had early to moderately advanced lung disease on CT, with well-preserved lung function. Patients with more advanced lung disease would be further affected by the uneven distribution of inhaled drugs. These findings match deposition studies in patients with CF, showing that the deposition pattern is more het - erogeneous in diseased lungs than in healthy lungs.127,128 In addition, it supports previous studies showing that penetration of inhaled drugs in deformed or partially obstructed airways is restricted. 128 These results suggest that upper lobes are more vulnerable to under-treatment and that this effect is stronger once structural damage is present. With our simulations, we showed that lower inhalation flows reduced extra-thoracic deposition, leading to higher AZLI concentrations in the small airways. This finding is consistent with previous studies showing that high flows lead to high extra-thoracic and upper airway drug deposition. 127 Using patient-specific airway modeling, we were able to study the impact of inhalation flow rate and inhaled volume on local airway drug con- centrations in the small airways. This information can be used to design smart nebulizers in such a way that adequate small airway concentrations can be obtained,25 or to define the required medication dose for a patient that, independent of the breathing pattern, results in sufficient drug delivery to critical areas of the lung. CFD offers a number of advantages that complement available techniques for study - ing aerosol deposition. Non-imaging techniques, e.g. pharmacokinetic methods, lack the ability to identify dose deposition into different zones of the lungs. 145 Scintigraphic methods do assess the deposition location of inhaled drugs, however, by dividing the lung into several large regions of interests. 146 3-helium MRI provides structural infor- mation and offers a quantification of ventilation down to the alveolar level, 147 however regional deposition of inhaled drugs cannot be derived from this technique. In contrast,
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 51
|
Modeling of antibiotic concentrations in CF patients CFD allows detailed information on aerosol deposition at specific anatomical sites to be determined. Another advantage is that CFD allows estimation of AZLI concentrations throughout the bronchial tree, data that are extremely difficult to obtain in vivo. To date, great emphasis has been given to sputum concentrations in clinical studies investigating inhaled antibiotics. It is highly likely that these concentrations are primarily reflective of central airway concentrations. The central airway concentrations found in this study were in the range of the fitted sputum concentrations from clinical studies, taking into account that published sputum samples were collected at later time-points compared to this study (data on file). As suggested by our findings, higher concentrations in central air- ways result in lower concentrations in small airways, challenging the validity of sputum samples as a useful indicator to explain the failure or success of inhaled antibiotics. We utilized CT-scans that were acquired as part of routine clinical care,148 allowing extraction of extra clinical relevant information without the need for additional radiation. Unlike other deposition study techniques, our model allowed us to study the impact of multiple variables, e.g. differences in particle size, on lung deposition within the same patient. CFD can therefore be used to predict lung deposition and effective dose of newly developed nebulizers. CFD opens up new pathways to further optimize inhalation therapies, even at a personalized level. Our modeling study has a number of limitations. To allow modeling of AZLI and estima- tion of concentrations, several assumptions were made. The first assumption was that the antibiotic concentration in the ASL is the most important determinant for effective killing of Pa.149-151 To estimate ASL concentrations we had to consider three different scenarios for ASL thickness. As ASL thickness cannot be measured in vivo, we used a number of ASL thicknesses in our model that covered the entire range found previously in in vitro data from CF bronchial epithelial cultures.140 We also did not take into account dissolution of the inhaled antibiotic in sputum that can cover the airway epithelia. 79 Although mucus layer thickness can vary between patients and throughout the bronchial tree, it is reason- able to assume that this ASL layer will be at least 3 μm (thinnest ASL of CF epithelia found in vitro). Thus, it is likely that the concentrations we computed are too optimistic. Areas covered by mucus, especially those in regions of the lung with severe disease, may have even lower antibiotic concentrations, potentially decreasing below MIC90. While ASL concentration is generally considered to be a reliable marker of alveolar antibiotic concentration,149-151 it is likely only an approximation as it relies on several as- sumptions. This model does not take into account drug uptake by alveolar macrophages as a measure of intracellular penetration in the lungs. 152 Especially in the chronically infected lung, macrophages may play a substantial role in the pharmacodynamics of anti-infective agents. This model also does not take into account binding of AZLI to spu- tum.152 Only unbound drug concentrations are considered to be microbiologically active. In a single study, it was observed that there was little binding of AZLI to CF sputum. 141
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 52
|
Chapter Therefore it is not likely that this effect has a substantial effect on our data. We did not account for mucociliary and cough clearance, which further reduces AZLI concentrations in the airways, and this occurs within minutes after inhalation.153 Our model assumes that the microbiological effect of a β-lactam antibiotic, such as AZLI, is best predicted using function of time above the MIC (T>MIC). 154,155 Unfortunately, the half-life of AZLI in the airways is not precisely known, but is thought to be approximately two hours in serum. Therefore, a prediction of efficacy of AZLI in the airways based on a single time-point immediately after inhalation is an approximation only. Thus, even though we calculated that the AZLI concentration was well above MIC 90 for most simulated conditions imme- diately after nebulization, concentrations may decrease well below MIC 90, especially in diseased areas, before a new dose of AZLI is nebulized. We estimated the concentration of AZLI that could be considered effective for kill - ing Pa strains based on previously reported data. However, the ideal AZLI concentration for effective killing of Pa in vivo is not well-defined, and varies largely between studies, with MIC90 values ranging between 32 and 128 μg/ml.15,42,141,156,157 Terms indicating the efficacy level of antibiotics, e.g. MIC and MIC90, have been used interchangeably in other studies, making comparison difficult.154,155 For AZLI, most susceptible bacteria are killed at concentrations 1 to 4-fold their MIC. However, antibiotic concentrations required for kill- ing Pa strains in biofilms are substantially higher than for killing Pa strains in planktonic growth. In an in vitro biofilm model, the time-dependent killing pattern of ceftazidime and imipenem in planktonic bacteria was changed to concentration-dependent killing for biofilm cells. Because of this, higher doses and longer treatment times with ceftazi - dime were required for the biofilm-growing Pa than for planktonic cells. While a concen- tration of 128xMIC was bactericidal for the wild-type strain (PAO1), a concentration of 2048xMIC was required for its β-lactamase overproducing mutant (PAΔDDh2Dh3).158 In the registration studies of AZLI, concentrations of more than 2048 μg/ml were required to achieve bactericidal killing of Pa in some cases. 42,47,157 This corresponds to an AZLI concentration of more than 16-fold the MIC90 referred to in this study. Studies on the ef- ficacy of AZLI mostly use a threshold of 10-fold MIC90, but without clear explanation.15,141 This threshold was also used for our analyses in this manuscript, but no claims can be made concerning its clinical significance. Airways with a diameter below 1-2 mm could not be reconstructed from CT data, and were added to the model using Phalen's description of the airway tree in infants, children and adolescents.125 These model data are derived from subjects without lung disease and were combined with the assumption of homogenous aerosol distribution in these small airways. In CF, it is well recognized that the small airways are progressively involved in early life lung disease. 5,6 Hence, it is likely that our model underestimates the het - erogeneity of aerosol deposition, with the assumption of normal structure of the small airways. Moreover, the small airways might be most prone to Pa infection122 and hence an
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 53
|
Chapter s1 TEXT. suPPlEmEnTAry mEThods. Chest computed tomography All routine chest CT scan sets were acquired using a 128-slice CT scanner (Somatom Definition Flash; Siemens, Erlangen, Germany). Thirty minutes prior to performing the CT scan, a lung function technician trained the study subjects by practicing the required spirometry manoeuvres in the supine position. Children were trained to obtain a breath- hold at maximal inspiration (total lung capacity, TLC) and maximal expiration (residual volume, RV) for 5–15 seconds. The inspiratory and expiratory slow vital capacities (SVC) achieved during the training were used as the reference values for the spirometric results during the CT scan. The reference SVCs were performed according to the ATS/ERS crite- ria.159 Breathing instructions during the CT scan were given by the same lung function technician. During the scan, the lung technician monitored in real time the inspired and expired volumes on the computer screen of the CT-compatible spirometer setup. When the patient reached the correct TLC (inspiratory scan) or RV (expiratory scan) breath hold level, the lung function technician signalled the CT-technician to start scanning. For the technician-guided technique the same breathing instructions were given during the CT scan; however, the inspired and expired volumes were not spirometrically measured. CT settings Tube voltages of 80kV (patients < 35kg) or 110kV (patients ≥ 35kg) were used with a 0.6s rotation time. Scanning was done from apex of the lung to base at 1.5 pitch and 6x2mm collimation. Images were reconstructed with a slice thickness ≤ 1.0mm, a slice increment ≤ 0.6mm and kernel B75f. For the inspiratory protocol, a modulating current was used (Siemens) with a reference tube current-time product of 20mAs for optimal image qual- ity. For expiratory CTs, a tube current fixed at 25mA with an effective tube current-time product of 10 mAs (the typical value for a 5-year-old child) was used, producing a lower radiation dose than the inspiratory protocol with sufficient image quality. Total radiation dose was in the order of 0.75 mSv for children below the age of 6 years and 1 mSv in older children. CT evaluation To quantify chest CT abnormalities, we used the validated CF-CT scoring system. 136 This scoring method evaluates the 5 lung lobes and the lingula as a sixth lobe for the fol- lowing components: 1) severity and extent of central and peripheral bronchiectasis; 2) severity and extent of central and peripheral airway wall thickening; 3) extent of central and peripheral mucus plugging; 4) extent of opacities (atelectasis, consolidation, ground glass pattern); 5) extent of cysts and bullae on inspiratory CTs and 6) the pattern and extent of trapped air on expiratory CTs. The maximal possible composite CT score is 207
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 54
|
Modeling of antibiotic concentrations in CF patients points. In the CF-CT scoring method, the CF-CT composite score is calculated by sum- ming the component scores per lobe. Instead of using the CF-CT composite score, the component scores per lobe were used for analysis. The lobar specific component scores were expressed as a percentage of the maximum possible component score per lobe. The component scores for bronchiectasis, airway wall thickening and air trapping were used for analysis. Prior to scoring, all CT scans were de-identified (Myrian®; Intrasense, Montpelier, France). Next, scans were scored in random order by an experienced observer, with more than 2 years' experience in scoring, who was blinded to clinical background. To assess inter-observer agreement, a second observer with 4 months scoring experience rescored all CT scans. Both observers were initially trained in CF-CT scoring using a standardized instruction module and training sets. Good intra- and inter-observer agreement was established on the training sets before scoring the study CT scans. To establish the intra- observer agreement, observer 1 rescored 25 random selected scans after 3 months. CF-CT scores of observer 1 were used for analysis.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 55
|
Chapter s1 Figure – Breathing profiles (high – Average – low) Inhalation part of breathing profiles. Low breathing profile: tidal volume of 6 ml/kg (228 ml) and respiration rate of 14 breaths per minute. Average breathing profile: tidal volume of 10 ml/kg (380 ml) and respiration rate of 18 breaths per minute. High breathing profile: tidal volume of 14 ml/kg (532 ml) and respiration rate of 22 breaths per minute. s2 Figure – Influence of increases in aerosol diameter and increases in airway surface liquid thickness on AZlI concentrations AZLI concentrations per lobe showing the influence of increases in aerosol diameter and increases in airway surface liquid (ASL) thickness on the AZLI concentrations. Data are presented as median (range). In each figure the AZLI concentrations are shown per separate lung lobe for the 3 different sizes of aerosol diameter. White bars represent the smallest aerosol diameter (2.9 μm), light grey bars represent the median aerosol diameter (3.18 μm) and dark grey bars represent the largest aero- sol diameter (4.35 µm). The separate parts of the figure represent the 3 different thicknesses of ASL. Part S2a shows the AZLI concentrations calculated for the thin ASL (3 μm), part S2b shows the AZLI concentrations calculated for the median ASL (5 μm) and part S2c shows the AZLI concentrations calculated for the thick ASL (7 μm). RUL = right upper lobe, RML = right middle lobe. RLL = right lower lobe, LUL = left upper lobe, LLL = left lower lobe. The larger the aerosol diameter and the thicker the ASL the lower were the AZLI concentrations. Note differences in the scale on the vertical axes between the 3 separate parts of the figure.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 56
|
Modeling of antibiotic concentrations in CF patients Table s1. P-values of pairwise comparison of AZLI concentrations between lobes, accompanying figure lobe Comparison with other lobes (p-values) rul rml rll lul lll RUL - 0.002 1.4E-09 0.156 2.7E-RML 0.002 - 2.7E-10 5.6E-06 2.7E-RLL 1.4E-09 2.7E-10 - 6.8E-11 5.3E-LUL 0.156 5.6E-06 6.8E-11 - 3.3E-LLL 2.7E-10 2.7E-10 5.3E-10 3.3E-13 - P-values of comparison between lobes in AZLI concentrations for the scenario of thick lining fluid with largest aerosol diameter. P-values in bold represent significant differences. RUL = right upper lobe, RML = right middle lobe. RLL = right lower lobe, LUL = left upper lobe, LLL = left lower lobe.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 57
|
Chapter ABsTrACT Background: Inhaled tobramycin is important in the treatment of Pseudomonas aerugi- nosa (Pa) infections in cystic fibrosis (CF). However, despite its use it fails to attenuate 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 inhalation (BID) to a once-daily (OD) inhaled double dose could improve treatment outcomes. Objectives: To predict local concentrations of nebulized tobramycin in the airways of patients with CF, delivered with the small airways-targeting Akita® or standard PARI-LC® Plus system, with different inspiratory flow profiles. methods: Computational fluid dynamic (CFD) methods were applied to patient-specific airway models reconstructed from chest computed tomography (CT) scans. The following BID and OD dosing regimens were evaluated: Akita® (150 and 300 mg) and PARI-LC® Plus (300 and 600 mg). Site-specific concentrations were calculated. results: Twelve CT-scans from patients aged 12-17 years (median=15.7) were selected. Small airways concentrations were 762-2999 μg/ml for BID and 1523-5997 μg/ml for the OD dosing regimen, which is well above the minimal inhibitory concentration (MIC) of wild type Pa strains. Importantly, the OD regimen appeared to be more suitable than the BID regimen against more resistant Pa strains and the inhibitory effects of sputum on tobramycin activity. Conclusions: CFD modeling showed that high concentrations of inhaled tobramycin are indeed delivered to the airways, with the Akita® being twice as efficient as the PARI-LC® system. Ultimately, the OD dosing regimen appears more effective against subpopula - tions with high MIC (i.e. more resistant strains).
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 58
|
Can we dose once daily? InTroduCTIon Inhaled tobramycin is important in the treatment of Pseudomonas aeruginosa (Pa) infec- tions in cystic fibrosis (CF) and is used to both eradicate early, and suppress chronic Pa infections.22 As tobramycin exhibits concentration-dependent bactericidal activity, the highest possible concentrations above the minimal inhibitory concentration (MIC) are re- quired in order to be optimally efficacious against Pa.160 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 grow- ing in biofilm. 161 Patients with CF are typically infected with multiple Pa morphotypes, each of which may vary in their susceptibility to antibiotics.162,Despite the use of current therapies to manage Pa infections, small airways disease in patients with CF continues to progress.5 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,164 concentrations in the small airways are as yet unknown and difficult to measure in vivo. Chronic treatment of Pa lung infection is currently defined as twice-daily tobramycin inhalation with the standard 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 to data on standard twice-daily inhalation, higher peak and lower trough levels were observed with the once-daily treatment regimen.165 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 computed tomography (CT) scans of CF patients who differ in disease severity, in combination with computational fluid dynamics (CFD) simulations, aerosol concentrations in the central and more distal airways can be com- puted. Furthermore, the relationship between airway morphology and airway concentra- tions can be assessed. This technique also yields similar information to SPECT CT (Single Photon Emission Computed Tomography) 130 and has recently been used to study the association between structural lung disease in CF and the deposition of Aztreonam ly- sine for inhalation (AZLI; inhaled antibiotic for Pa treatment).166 Unlike tobramycin, using concentrations above the MIC does not improve the bactericidal activity of AZLI (i.e. not concentration-dependent killing) and simulations were run for a different nebulizer (i.e.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 59
|
Chapter eFlow). Therefore, the results of this study cannot be simply extrapolated to tobramycin nebulization. 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 computational fluid dynamic (CFD) simulations. 166 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 which we selected the 6 youngest and 6 oldest airway models for this study. All scans were anonymized and scored in random order by 2 experienced observers, whereby the validated CF-CT scoring system 136 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 are 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. 166 Briefly, the deposition of in - haled 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. Approval for this retrospective study was obtained from the Institutional Review Board of the Erasmus Medical Center 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 three-dimensional 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 pediatric
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 60
|
Can we dose once daily? population. The upper airway model was scaled twice, once for the 6 youngest patients and once for the 6 oldest patients. These models were connected to the mouthpiece of the PARI-LC® Sprint nebulizer, which was further connected to the patient-specific lower airway model (Figure S1). reconstruction of three-dimensional 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. 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 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 5-8 seconds and 1.0-1.6L, respectively. The breathing profile had an inspiration-expiration ratio of 1:1.5 and was sinusoidal in shape. Tobramycin nebulization commenced immedi- ately 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/second and a mass median aerosol diameter (MMAD) of 3.6 and geometric standard deviation (GSD) of 2.0 µm were used for inhalation of Bramitob by the Akita®.Loading doses of 150 mg and 300 mg tobramycin were used in simulations for the Akita® system. Delivered doses are shown in Table 1 and were based on an in vitro study Table 1 – Doses used for simulations Akita® PArI-lC® Plus Loading dose BID 150 Loading dose OD 300 Delivered dose BID 26.95 96.Delivered dose OD 53.9 192.Loading doses and delivered doses used for the simulations. Doses are in milligrams. These doses are based on an in vitro study which aimed to calculate the required delivered dose with the Akita® nebu- lizer to obtain an equivalent lung dose as the PARI-LC® Plus nebulizer. Due to the efficiency of the Akita®, the delivered dose for this system was set to be much lower to obtain an equivalent lung dose to the PARI-LC® Plus. Both nebulizers in this study were filled with 300 mg tobramycin.12, 13 Thus for the Akita® these doses belong to the once-daily regimen in our study and for the PARI-LC® Plus these doses belong to the twice-daily regimen. BID = twice-daily; OD = once-daily.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 61
|
Chapter investigating the required delivered doses of different nebulizers to obtain an equivalent lung dose to the PARI-LC® Plus nebulizer.167,168 Due 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. 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 nebuliza - tion was continuous during both inspiration and expiration. Furthermore, these profiles were generated using the age and length of each patient at time of the CT scan, and the reference formula developed by Zapletal et al.169 Specifically, the mean, upper and lower limits of this formula were used to generate a mean, high and low tidal volume, respec - tively. 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 com - pressor)170 and largest MMAD (4.93 µm; TOBI® nebulized with PARI-LC® Plus combined with DevilBiss PulmoAide compressor). 171 Additionally, a GSD of 2.3 was used in these simulations.Loading doses of 300 mg and 600 mg 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 exceeding 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 smaller than 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. 125 Phalen's data were obtained from subjects without lung disease and describes airway dimensions up to the sixteenth generation. Once the aerosol entered a lobe in the Phalen section of the model, it was as- sumed that it would be distributed homogeneously. Regional tobramycin concentrations
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 62
|
Can we dose once daily? 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, 7 μm).140 Tobramycin concentrations were calculated for the following variables: the Akita® (1 MMAD, 1 breathing profile) or Pari LC Plus® (2 MMADs, 3 tidal volumes) systems, ELF thickness (3, 5, 7 μm) and the dosing regimen (once- and twice-daily dose). Results are described for the median ELF thickness of 5 μm unless otherwise indicated. Statistical analysis Patient characteristics and tobramycin concentrations were summarized using descrip- tive statistics and all data are presented as the median (interquartile range). 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, FEV 1 and FEF 75 % pred. The advantage of using mixed-effects models is that they account for repeated measure- ments on the same patients. Descriptive statistics and intraclass correlation coefficients (ICC) 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 in- cluded from two different time points, two 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 % predicted, was highly variable. 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 77255 μg/ml when dosed twice daily and 154511 μg/ml when dosed once daily. For the PARI-LC® Plus, median tobramycin concentrations in the large airways were
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 63
|
Chapter 84316-94957 μg/ml when dosed twice daily and 168633-189916 μg/ml when dosed once daily. Figure 1 shows tobramycin concentrations in the small airways calculated for the median ELF. Simulations with the Akita® resulted in tobramycin concentrations of 1770 μg/ml when dosed twice daily and 3541 μg/ml when dosed once daily. For the PARI- LC® Plus, tobramycin concentrations were 1066-1800 µg/ml when dosed twice daily and 2133-3598 μg/ml when dosed once daily. Figure 1 also shows how tidal volume Table 2 – Baseline characteristics Value IQ range Male 33.3% Age 15.7 (13.9-17.1) Total airway disease score (% of max score) 7.3 (4.1-11.1) FEV1%pred 92.6 (86.4-115.9) FEF75%pred 54.7 (35.2-71.5) Median (interquartile (IQ) range). FEV1%pred = percent predicted forced expiratory volume in one sec - ond; FEF75%pred = percent predicted forced expiratory flow at 75% of forced vital capacity. Figure 1 – Small airways tobramycin concentrations Small airways tobramycin concentrations calculated for the median epithelial lining fluid of 5 µm. Concentrations after nebulization of the once-daily dose (boxplot on the left) are higher than con- centrations after nebulization of the twice-daily dose (boxplot on the right). Concentrations for the once-daily dose range between approximately 2000 and 4000 μg/ml. Concentrations for the twice- daily dose range between approximately 1000 and 2000 μg/ml. Highest concentrations are ob- tained with the Akita® and PARI-LC® Plus nebulized with a lower tidal volume and small MMAD of 3.4 µm. For the PARI-LC® Plus, higher concentrations are obtained with the small MMAD of 3.4 µm than the large MMAD of 4.93 μm. MMAD = mass median aerosol diameter.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 64
|
Chapter 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. Small airways concentrations for all ELF thicknesses are shown in Figure S2 of the Supplementary data. Significantly lower tobramycin correlations were observed with the PARI-LC® Plus nebulizer than with the Akita® nebulizer, with the exception of inhalation at a low breath- ing tidal volume and the smallest MMAD (3.4 μm) (Table 3). For example, if two patients with the same TAD score, FEV1 % predicted and FEF 75 % predicted 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 µg/ml (once- and twice-daily, respectively) compared to the Akita®. Much smaller and reverse associations were seen for FEF75 % predicted and the TAD score. dIsCussIon In this study, CFD was used to compute local inhaled tobramycin concentrations through- out 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 once-daily dose resulted in higher tobramycin concentrations in the small airways compared to the twice-daily dose (1523-5997 μg/ml versus 762-2999 μg/ml, respectively). This result is promising due to the concentration- dependent bactericidal efficacy of tobramycin; where the higher the concentration in a specific region, the greater the reduction in bacterial density.However, whether the computed concentrations are sufficient for effective killing throughout the lung remains questionable, as the concentration at which effective kill - ing is obtained is as yet unknown. Due to the substantial heterogeneity of phenotypes and genotypes of Pa within a single patient, a range of MIC values exist throughout the lung.126 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 micro-organisms 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 mucus 93,94 and due to biofilm formation by Pa.174 To overcome the effect of sputum binding on tobra - mycin, it is generally assumed that local concentrations need to be 10 to 25-fold above the MIC. Wild type 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 wild types have a MIC of 2 μg/ml or lower.175,176
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 65
|
Can we dose once daily? Hence for wildtype Pa (MIC2 μg/ml), effective killing can be easily achieved as concentra - tions 381-1500x MIC 2 μg/ml were observed in the small airways for the twice-daily dose and 762-2999x MIC 2 μg/ml for the once-daily dose. However, for Pa strains with in vitro MICs above 2 μg/ml (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 μg/ml (MIC512 μg/ml), only the once-daily dose calculated for a thin lining fluid resulted in concentrations 10-fold greater than the MIC (10x MIC 512 μg/ml). For these simulations, all patients received small airways concentrations above 10x MIC512 μg/ml 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 (Figures 2 and S3). 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.96 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. To overcome the antagonistic activity of mucus in all patients, including those with the worst-case sputum composi- tion, peak tobramycin concentrations had to be 100-times the MIC to ensure killing of planktonic Pa,177 or even 100-1000-fold greater for Pa growing in biofilm.161 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 com- mence 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 approximately 2 hours 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 (approximately 5 hours) than the effect in vitro,178 as longer half-lives in- crease the duration of this effect.179 The half-life of tobramycin measured in sputum was approximately 2 hours post-nebulization of 80 mg tobramycin.180
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 66
|
Can we dose once daily? 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 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. Additionally, studies in intravenous tobra- mycin showed that a longer clearance period of systemically absorbed tobramycin is associated with a better safety profile,181,182 suggesting that a once-daily, double dose of inhaled tobramycin is safe.165 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 AZLI in relation to structural lung disease in CF patients showed that the more diseased lobes received lower levels of the inhaled antibiotic,166 and this is important for inhaled tobramycin due to its low rate of systemic absorption (9-17.5%).12 This means that hypoventilated lung areas are under- treated 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),126 a clinical response in these hypoventilated areas is likely to occur following systemic absorption via the bronchial circulation. The limitations of CFD modeling have been previously described. 166 Firstly, in order 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 below 1-2 mm, and breathing profiles. An exception to this limitation would be the breathing profiles for the Akita®, which were based on the FEV 1 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 modeling 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 We demonstrated that high concentrations of inhaled tobramycin are indeed delivered to the lung, 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
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 67
|
Chapter Figure s2 – small airway tobramycin concentrations calculated for different lining fluid thickness Thin lining fluid (3 µm) Small airways tobramycin concentrations calculated for the thin epithelial lining fluid of 3 µm. Con- centrations after nebulization of the once-daily dose (boxplot on the left) are higher than concentra- tions after nebulization of the twice-daily dose (boxplot on the right). Concentrations for the once- daily dose range between approximately 3500 and 6000 μg/ml. Concentrations for the twice-daily dose range between approximately 1750 and 3000 μg/ml. Highest concentrations are obtained with the Akita® and PARI-LC® Plus nebulized with a lower tidal volume and small MMAD of 3.4 µm. For the PARI-LC® Plus, higher concentrations are obtained with the small MMAD of 3.4 µm than the large MMAD of 4.93 μm. MMAD = mass median aerosol diameter. Median lining fluid (5 µm) Small airways tobramycin concentrations calculated for the median epithelial lining fluid of 5 µm. Concentrations after nebulization of the once-daily dose (boxplot on the left) are higher than con- centrations after nebulization of the twice-daily dose (boxplot on the right). Concentrations for the once-daily dose range between approximately 2000 and 4000 μg/ml. Concentrations for the twice- daily dose range between approximately 1000 and 2000 μg/ml. Highest concentrations are ob-
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 68
|
Can we dose once daily? tained with the Akita® and PARI-LC® Plus nebulized with a lower tidal volume and small MMAD of 3.4 µm. For the PARI-LC® Plus, higher concentrations are obtained with the small MMAD of 3.4 µm than the large MMAD of 4.93 μm. MMAD = mass median aerosol diameter. Thick lining fluid (7 µm) Small airways tobramycin concentrations calculated for the thick epithelial lining fluid of 7 µm. Concentrations after nebulization of the once-daily dose (boxplot on the left) are higher than con- centrations after nebulization of the twice-daily dose (boxplot on the right). Concentrations for the once-daily dose range between approximately 1500 and 2500 μg/ml. Concentrations for the twice- daily dose range between approximately 750 and 1300 μg/ml. Highest concentrations are obtained with the Akita® and PARI-LC® Plus nebulized with a lower tidal volume and small MMAD of 3.4 µm. For the PARI-LC® Plus, higher concentrations are obtained with the small MMAD of 3.4 µm than the large MMAD of 4.93 μm. MMAD = mass median aerosol diameter.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 69
|
Chapter Nebuliser and tidal volume used for simulation Pari LC Plus with higher tidal volume Pari LC Plus with lower tidal volume Pari LC Plus with mean tidal volume Akita 4.93 mcm (Pari LC Plus) 3.4 mcm (Pari LC Plus) 3.6 mcm (Akita) MMAD MIC: 5120 mcg/ml Page Figure s3 – Boxplot small airways concentrations once-daily dosing calculated for a thin lining fl uid related to 10xmIC512 μg/ml Small airways tobramycin concentrations after nebulization of the once-daily dose calculated for the thin epithelial lining fl uid of 3 µm. All patients received small airways concentrations above 10xMIC512μg/ml with the Akita® and PARI-LC® Plus when nebulized with a low tidal volume and small MMAD of 3.4 μm. None of the patients received concentrations above this level with the Pari LC Plus nebulized with a mean or high tidal volume and large MMAD of 4.93 μm. Only part of the patients received concentrations above the threshold in the other simulated situations for the PARI-LC® Plus. The fi gure also shows the direct infl uence of tidal volume and MMAD on the concentrations in the small airways for the PARI-LC® Plus. Lower tobramycin concentrations in the small airways are observed with higher tidal volumes and the largest MMAD. MMAD = mass median aerosol diameter.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 70
|
Chapter ABsTrACT Background: Small airways disease (SAD) progresses with age in most cystic fibrosis (CF) patients in spite of maintenance therapy including inhaled dornase alfa. In a clinical study, efficient targeting of dornase alfa to the small airways using a smart nebulizer improved SAD in stable patients with CF on maintenance dornase alfa therapy, but not in patients who were admitted for an exacerbation. We hypothesized that higher than normal doses are required for patients with more severe lung disease and that suboptimal dornase alfa concentrations are obtained in the small airways with conventional nebulizers. methods: Computational fluid dynamic (CFD) modeling was applied to patient-specific airway models reconstructed from chest CT-scans (FLUIDDA nv, Kontich, Belgium). Disease severity was scored using CT scoring systems. The following settings were used for simulations: nebulization of 2.5 mg dornase alfa, nebulizer specific aerosol characteristics, and epithelial lining fluid thickness of 5 µm. CFD was performed for the smart Akita® nebulizer targeted to the large airways (Akita large) and to the small airways (Akitasmall), and for PARI-LC® Plus or eFlow® nebulizer. Concentrations are expressed relative to an effective dornase alfa concentration of 2.9 µg/ml. Results are reported as median [interquartile range]. results: 40 CT-scans (35% male) were selected, age 10.9 [8.7-15.2] years, FEV 1 94.3 [81.7-102.5] %pred. Large airways concentrations for the Akita large and Akita small were 15 and 12-13 times higher than for the PARI-LC® Plus and eFlow® (8677, 7225, 588 and 571 μg/ml, respectively). Small airways concentrations were 27-30 and 17-20 times higher (p <0.001) than for the other nebulizers (Akita large: 173.4 [166-196], Akita small: 111.4 [103-123], PARI-LC® Plus: 6.5 [6-7], eFlow®: 5.6 [5-6] µg/ml). 4.3% and 8.0% of the lung lobes received dornase alfa concentrations below 2.9 µg/ml with the PARI-LC® Plus and eFlow®, respectively. Conclusions: For a breath actuated nebulizer dornase alfa concentrations in small air- ways were well above the minimal effective concentration as computed by CFD. For con- ventional nebulizers, concentrations were around or below the effective concentration especially in diseased areas. CFD simulations confirm observations of a clinical study where higher doses were more effective in reversing small airways obstruction.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 71
|
Modeling of dornase alfa concentrations in CF patients InTroduCTIon Cystic fibrosis (CF) lung disease is characterized by chronic early lung infection and inflammation. As part of this process, deoxyribonucleic acid (DNA) is released from inflammatory cells and bacteria that lyse within the airways. DNA significantly increases the viscoelasticity of mucus causing progressive obstruction due to mucus impaction in the airways, including the small airways. 20,183 This small airways disease (SAD) plays a major role in the pathophysiology of progressive CF lung disease.Inhaled dornase alfa was developed in the early Nineties to reduce mucus viscosity in CF by enzymatic cutting of free DNA. 19 It was shown that daily nebulization of 2.5 mg dornase alfa reduced obstruction as measured by spirometry and multiple breath washout, reduced the number of exacerbations, and reduced the rate of lung function decline.19,49 No clear dose response relation was observed between a once daily admin- istration of 2.5 mg dornase alfa and a twice daily administration. 184 For this reason, the recommended dose for maintenance therapy with dornase alfa is once daily 2.5 mg, nebulized with an approved jet nebulizer/compressor combination or the Pari eFlow Rapid® nebulizer system (further referred to as eFlow).20 Daily treatment of dornase alfa has become standard treatment for CF patients of 6 years and above. However, despite this treatment residual SAD progresses with age in most CF patients. 9 For this reason it was questioned whether high enough doses could ever be deposited in the small airways using the conventional nebulizers which are operated while the patient is using tidal volume breathing.185 Furthermore, it is well recognized that the nebulizers used in the pivotal trials for dornase alfa are quite inefficient relative to today's state of the art smart nebulizers. In addition, nebulizers like eFlow used in today's clinical practice have different characteristics than those used in the pivotal trials. To investigate whether more efficient delivery of dornase alfa to the small airways would reverse SAD in stable CF patients a randomized controlled study was executed in a small number of patients by Bakker et al.25 In stable CF patients with SAD and who were on maintenance dornase alfa therapy, the use of a conventional nebulizer was switched to a smart Akita® nebulizer (Vectura, UK). Patients were randomized to the Akita set to target the large airways or to the Akita set to target the small airways. This smart nebulizer controls and guides the flow and volume of inhalation of aerosol and reduces loss of aerosol into the environment by limiting nebulization to the inspiratory phase. As a result lung deposition is improved to 70% of the loading dose compared to 10-15% using conventional nebulizers. Therefore, the total lung dose during this study was estimated to be 3-5 times higher compared to conventional nebulizers. The switch to the Akita® nebulizer resulted in substantial im - provement of small airways patency.25 The effect was larger for the small airways group. In a similar study by Bakker et al.,185 it was investigated whether efficient targeting of dornase alfa to the small airways also would improve SAD in CF patients with a respira-
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 72
|
Chapter tory tract exacerbation. CF patients hospitalized for an exacerbation were randomized to small or large airway deposition of dornase alfa using the Akita nebulizer. In this study, for safety reasons, half the dose of dornase alfa was given (1.25 mg) which was estimated to result in a lung dose that would be 1.5-2 times higher than conventional nebuliz - ers. The switch to the Akita nebulizer in these unstable CF patients hospitalized for an exacerbation did not result in more effective treatment of small airways obstruction than other nebulizers in contrast to the study in stable patients. This may have been caused by the lower dose administered in this second trial. As sputum production and central deposition is increased during respiratory tract exacerbations this lower dose may have resulted in under-dosing of the small airways.Based on the above described two studies, we raised the question whether efficient targeting of dornase alfa to the small airways using a smart nebulizer would result in suf- ficiently high concentrations in the small airways. We hypothesized that, even in stable patients with SAD, dornase alfa concentrations are not high enough throughout the lung. To answer this question we used computational fluid dynamic (CFD) simulations to estimate aerosol concentrations in large and small airways in relation to disease severity. Using CFD, we computed the percentage of airways that received suboptimal dornase alfa concentrations when dornase alfa was nebulized using either a conventional PARI- LC® Plus nebulizer, an eFlow nebulizer, or the smart Akita nebulizer set to target the large or small airways. mEThods reconstruction of three-dimensional airway models and lung lobes We used 40 patient-specific 3D models of the airways and lung lobes that were reconstructed previously and used for CFD simulations. 166 These airway models were reconstructed from spirometry controlled inhalation- and expiration CT scans of patients with CF aged 6-18 years, acquired as part of routine clinical care. Reconstruction of the airway models was performed by FLUIDDA nv (Kontich, Belgium) as has been described in detail in a previous paper.Briefly, a 3D model of the intra-thoracic region down to the level of airways with a diam- eter of 1-2 mm was reconstructed from the inspiratory scan, using automatic segmenta- tion. Missing branches were manually added and incorrect branches manually deleted, as necessary. The upper airways were not included in the CT-scans. Therefore, an upper airway model from an average adult was scaled down to match the average tracheal diameter for the population. This model was connected to the nebulizer specific mouth- pieces. For the eFlow the mouthpiece was obtained using reverse engineering, for the
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 73
|
Modeling of dornase alfa concentrations in CF patients PARI-LC Plus, and Akita - PARI-LC Sprint combination a digital CAD model of the PARI-LC Sprint mouthpiece was used as the mouthpieces of the PARI-LC Plus and PARI-LC Sprint are equivalent. The patient-specific lower airway model was connected to the nebulizer mouthpiece/upper airway model (Supplementary figure S1). Using a semi-automated image analysis tool, the patient-specific lung lobe volumes were extracted from inspiratory and expiratory CT scans. This tool identifies the fissures separating the lobes. The volume change from expiration to inspiration was used for each lobe to compute the distribution of inhaled aerosol to that specific lobe. scoring of Chest CT abnormalities To quantify chest CT abnormalities, scans were anonymized and scored in random order using the validated CF-CT136 and PRAGMA-CF scoring systems.186 With the CF-CT scoring system, on the inspiratory scan, the five lung lobes and lingula were evaluated in the central and peripheral lung regions for severity and extent of the following components: bronchiectasis, airway wall thickening, mucous plugging and opacities (atelectasis, con - solidation, ground glass pattern).187 A total airway disease score (TAD) score was calculated as the sum of bronchiectasis, airway wall thickening and mucous plugging and expressed as a percentage of the maximum score (0-100%) (CFCT-TAD). PRAGMA-CF uses a grid overlaying 10 equally spaced axial CT slices between apex and base. For each grid cell the following CT components are scored on inspiratory scans in hierarchical order (high- est to lowest priority): 1. bronchiectasis; 2. mucous plugging; 3. airway wall thickening; 4. atelectasis or 5. normal lung. Trapped air is assessed with a similar methodology on the expiratory scans.186 For each component the volume is then computed and expressed as a fraction of total lung volume. As for the CF-CT scoring system, a TAD score was calculated as the sum of bronchiectasis, mucous plugging and airway wall thickening (PRAGMA-TAD).CFD simulations CFD simulations were performed by FLUIDDA nv (Kontich, Belgium). Deposition of in- haled dornase alfa was simulated for the Portaneb compressor with the PARI-LC Plus nebulizer, for the Akita-Jet with the PARI-LC Sprint nebulizer set to target small airways and set to target large airways, and for the eFlow nebulizer. Breathing profile and aerosol characteristics for Akita For dornase alfa deposition with the Akita we used the settings simulating the conditions used in the clinical trial involving stable patients with CF. 25 Inhalation time and volume for both settings (large and small airways) were based on the individual value for Forced Expiratory Volume in 1 second (FEV 1) at the time of the CT-scan for each individual patient. To generate the two different lung deposition patterns, three characteristics of the nebulizer were adjusted: particle size, breathing pattern and timing of aerosol bolus.
|
Erasmus
|
Erasmus
|
Erasmus_BosAukje_2016-11-23_161123_Bos-Aukje-Catharina.pdf
| 74
|
Chapter For the Akita targeting the small airways (Akita Small), the inhalation time and volume per breath varied between 3.8-7.5 seconds and 0.75-1.5 L, respectively. The breathing profile had an inspiration-expiration ratio of 1:1.5. The inspiration part was stepwise and expiration part sinusoidal (Supplementary figure S2). Dornase alfa nebulization com - menced immediately after inspiration. The inhalation phase consisted a period of the delivery of the dornase alfa aerosol followed by 1 second of fresh air. This air bolus helps to transport the aerosol deeper down in the small airways. The inhalation flow rate was fixed at 200 ml/second and a mass median aerosol diameter (MMAD) of 4.2 and GSD of 1.6 μm were used.For the Akita targeting the large airways (Akita large) simulating deposition of dornase alfa delivered by a conventional nebulizer, the inhalation time and volume per breath varied between 3.8-5.3 seconds and 0.4-0.7 L, respectively. Again, the inspiration- expiration ratio was 1:1.5, with a stepwise inspiration and sinusoidal expiration (Supple- mentary figure S3). In contrast to Akita small, dornase alfa nebulization did not commence immediately after the start of inspiration. Instead, the inspiration started with a bolus of air without aerosol. Depending on the inhalation time the bolus of fresh air continued for 0.5 to 2.5 seconds after which nebulization started. Nebulization continued for 2.5 seconds regardless of inhalation time, after which, again, a bolus of air without aerosol was inhaled during the last 0.3 seconds of the inhalation. As inhalation time per breath varied between 3.3 and 5.3 seconds, nebulization started 0.5 to 2.5 seconds after start of inspiration. The inspiration cycle was followed by a breath-hold of 3.0 seconds. During the inspiration cycle, inhalation flow rate was 200 ml/second when nebulization was off and 60 ml/second when nebulization was on (Supplementary figure S3). A MMAD of 6.0 and GSD of 1.6 μm were used.A loading dose of 2.5 mg dornase alfa was used in simulations for the Akita system. On average 97.18% of the loading dose is delivered to the lungs for the Akita targeting the small airways and 92.33% for the Akita targeting the large airways. This results in an ex- pected delivered dose of 2.43 mg for AkitaSmall (2.5 mg * 97.18% = 2.43 mg) and expected delivered dose of 2.31 mg for Akita large (2.5 mg * 92.33% = 2.31 mg). These delivered doses were used for the simulations. Both inspiration and expiration were modelled. Breathing profile and aerosol characteristics PARI-LC Plus and eFlow The sinusoidal breathing profiles for the simulations of dornase alfa deposition with the PARI-LC Plus and eFlow nebulizers had an inspiration-expiration ratio of 1:1.5, where dornase alfa nebulization was continuous during both inspiration and expiration. Fur- thermore, these profiles were generated using the specific age and height of each patient at the time of the CT-scan and the reference formula developed by Zapletal et al. (mean value).169 The PARI-LC Plus generates an aerosol with a MMAD of 4.6 and GSD of 2.14 μm.188 The eFlow generates an aerosol with a MMAD of 4.8 and GSD of 1.8 μm.188
|
Erasmus
|
End of preview. Expand
in Data Studio
We use OAI-PMH to download metadata for rights-free non-embargoed medical PhD theses, followed by PDF downloading and document parsing.
For now we have theses from Dutch universities.
- Downloads last month
- 15