The effect of formulation composition and adjuvant type on difenoconazole dislodgeable foliar residue.

Rigorous risk assessments for those exposed to pesticides are carried out to satisfy crop protection regulatory requirements. Non-dietary risk assessments involve estimating the amount of residue which can be transferred from plant foliage to the skin or clothes, known as dislodgeable foliar residues (DFRs). DFR data are less available than crop residue data as studies are costly and limited by seasonality. European regulatory authorities are reticent to allow extrapolation of study data to different scenarios as the contributory factors have hitherto been poorly identified. This study is the first to use a new laboratory DFR method to investigate how one such factor, pesticide formulation, may affect DFR on a variety of crops. The study used the active substance difenoconazole as both an emulsifiable concentrate (EC 10%) and a wettable powder (WP 10%) with and without adjuvants (Tween 20 and organophosphate tris(2-ethylhexyl)phosphate TEHP) on tomato, French bean and oilseed rape. A comparable DFR% was retained from the WP and EC formulation on most crops except for tomato, where lower DFR% was retained in the case of WP (39 ± 4.7%) compared to EC (60 ± 1.2%). No significant effect of adjuvant addition was observed for either formulation except when mixing TEHP (0.1% w/v) to the EC 10% on French bean, resulting in 8% DFR reduction compared to the EC formulation alone. This research demonstrates the value of a unique DFR laboratory technique in investigating the importance of the formulation and in-tank adjuvants as factors that affect DFR.

Stimuli-Responsive Polymers for Engineered Emulsions.

Emulsions are complex. Dispersing two immiscible phases, thus expanding an interface, requires effort to achieve and the resultant dispersion is thermodynamically unstable, driving the system toward coalescence. Furthermore, physical instabilities, including creaming, arise due to presence of dispersed droplets of different densities to a continuous phase. Emulsions allow the formulation of oils, can act as vehicles to solubilize both hydrophilic and lipophilic molecules, and can be tailored to desirable rheological profiles, including "gel-like" behavior and shear thinning. The usefulness of emulsions can be further expanded by imparting stimuli-responsive or "smart" behaviors by inclusion of a stimuli-responsive emulsifier, polymer or surfactant. This enables manipulation like gelation, breaking, or aggregation, by external triggers such as pH, temperature, or salt concentration changes. This platform generates functional materials for pharmaceuticals, cosmetics, oil recovery, and colloid engineering, combining both smart behaviors and intrinsic benefit of emulsions. However, with increased functionality comes greater complexity. This review focuses on the use of stimuli-responsive polymers for the generation of smart emulsions, motivated by the great adaptability of polymers for this application and their efficacy as steric stabilizers. Stimuli-responsive emulsions are described according to the trigger used to provide the reader with an overview of progress in this field.

Microstructural insight into inhalation powder blends through correlative multi-scale X-ray computed tomography.

Dry powder inhalers (DPI) are important for topical drug delivery to the lungs, but characterising the pre-aerosolised powder microstructure is a key initial step in understanding the post-aerosolised blend performance. In this work, we characterise the pre-aerosolised 3D microstructure of an inhalation blend using correlative multi-scale X-ray Computed Tomography (XCT), identifying lactose and drug-rich phases at multiple length scales on the same sample. The drug-rich phase distribution across the sample is shown to be homogeneous on a bulk scale but heterogeneous on a particulate scale, with individual clusters containing different amounts of drug-rich phase, and different parts of a carrier particle coated with different amounts of drug-rich phase. Simple scalings of the drug-rich phase thickness with carrier particle size are used to derive the drug-proportion to carrier particle size relationship. This work opens new doors to micro-structural assessment of inhalation powders that could be invaluable for bioequivalence assessment of dry powder inhalers.

Predicting the Strength of Cohesive and Adhesive Interparticle Interactions for Dry Powder Inhalation Blends of Terbutaline Sulfate with α-Lactose Monohydrate.

Grid-based systematic search methods are used to investigate molecule-molecule, molecule-surface, and surface-surface contributions to interparticle interactions in order to identify the crystal faces that most strongly affect particle behavior during powder blend formulation and delivery processes. The model system comprises terbutaline sulfate (TBS) as an active pharmaceutical ingredient (API) and α-form lactose monohydrate (LMH). A combination of systematic molecular modeling and X-ray computed tomography (XCT) is used to determine not only the adhesive and cohesive interparticle energies but, also the agglomeration behavior during manufacturing and de-agglomeration behavior during delivery after inhalation. This is achieved through a detailed examination of the balance between the adhesive and cohesive energies with the XCT results confirming the blend segregation tendencies, through the particle-particle de-agglomeration process. The results reveal that the cohesive interaction energies of TBS-TBS are higher than the adhesive energies between TBS and LMH, but that the cohesive energies of LMH-LMH are the smallest between molecule and molecule, molecule and surface, and surface and surface. This shows how systematic grid-search molecular modeling along with XCT can guide the digital formulation design of inhalation powders in order to achieve optimum aerosolization and efficacy for inhaled medicines. This will lead to faster pharmaceutical design with less variability, higher quality, and enhanced performance.

A new laboratory method to study the impact of leaf texture on pesticide dislodgeable foliar residues (DFR).

Pesticides are vital in meeting the challenge of feeding the rapidly increasing world population. However, it is crucial that they are used in a way that does not compromise the safety of humans or the environment. Non-dietary worker risk assessments consider the amount of residue which can be transferred from plant foliage to the skin or clothes, known as dislodgeable foliar residues (DFRs). DFR data scarcity due to the costly and seasonal characteristics of DFR studies is an obstacle to the extrapolation of DFR data to different crops/leaves. This paper validates a new proof-of-concept technique to investigate factors that may affect DFR (leaf texture) using the fungicide difenoconazole EC 10% as an example on various leaves (i.e., French bean, soybean, tomato, oilseed rape, and wheat). DFR was the lowest in the case of oilseed rape (31.0 ± 3.4%) and the highest in French beans (82.0 ± 2.9%). This significant difference in DFR in the findings of this study sheds light on the importance of the leaf surface as a major factor affecting DFR and supports the application of the laboratory method for more extensive data generation. More data generation would enable the extrapolation saving money and resources.

Profiling alveolar macrophage responses to inhaled compounds using in vitro high content image analysis.

One of the main hurdles in the development of new inhaled medicines is the frequent observation of foamy macrophage (FM) responses in non-clinical studies in experimental animals, which raises safety concerns and hinders progress into clinical trials. We have investigated the potential of a novel multi-parameter high content image analysis (HCIA) assay as an in vitro safety screening tool to predict drug induced FM. Rat (NR8383) and human U937-derived alveolar macrophages were exposed in vitro to a panel of model compounds with different biological activity, including inhaled bronchodilators, inhaled corticosteroids (ICS), phospholipidosis inducers and proapoptotic agents. An HCIA was utilized to produce drug-induced cell response profiles based on individual cell health, morphology and lipid content parameters. The profiles of both rat and human macrophage cell lines differentiated between cell responses to marketed inhaled drugs and compounds known to induce phospholipidosis and apoptosis. Hierarchical clustering of the aggregated data allowed identification of distinct cell profiles in response to exposure to phospholipidosis and apoptosis inducers. Additionally, in NR8383 cell responses formed two distinct clusters, associated with increased vacuolation with or without lipid accumulation. U937 cells presented a similar trend but appeared less sensitive to drug exposure and presented a narrower range of responses. These results indicate that our multi-parameter HCIA assay is suitable to generate characteristic drug-induced macrophage response profiles, thus enabling differentiation of foamy macrophage phenotypes associated with phospholipidosis and apoptosis. This approach shows great potential as pre-clinical in vitro screening tool for safety assessment of candidate inhaled medicines.

Development of a New Dislodgeable Foliar Residue Analytical Laboratory Method for Pesticides.

The dislodgeable foliar residue (DFR) is the amount of pesticide that exists on foliage after the pesticide has dried and which could dislodge to the skin or clothes of workers and is a key parameter for non-dietary risk assessments required to demonstrate safe use for pesticide registration. DFR data in the literature are described as insufficiently reliable, limited, and encompasses considerable statistical uncertainties. The purpose of this article is to describe a newly developed laboratory method for the quantification of DFR with an illustrative example. The laboratory method reflected available field DFR methodology but involved controlled application of droplets to leaves and validation of the wash-off process used to remove the residue from the leaf surface before the analytical quantification. A very high level of accuracy (99.7-102.1%) and precision (±1.5%) was achieved. Residue data generated from the illustrated application of the method showed a robust normal distribution, unlike field studies. The method is deemed to be controllable, cost-efficient, and time-saving, taking hours rather than days. This enables the generation of more data to allow extrapolation between the generated data by investigating multiple factors that may influence DFR. An improved understanding of DFR could save time, money, and resources.

High Content Image Analysis as a Tool to Morphologically Distinguish Macrophage Activation and Determine Its Importance for Foamy Alveolar Macrophage Responses.

Introduction: Lung diseases are an increasing global health burden affecting millions of people worldwide. Only a few new inhaled medicines have reached the market in the last 30 years, in part due to foamy alveolar macrophage (FAM) responses observed in pre-clinical rat studies. The induction mechanism and signaling pathways involved in the development of highly vacuolated 'foamy' phenotype is not known. Furthermore, it has not been determined if these observations are adaptive or adverse responses. Aim: To determine if high content image analysis techniques can distinguish between alveolar macrophage activation (LPS/IFN-γ activated and IL-4 activated macrophages) and if this could be applied to understanding the generation of 'foamy' macrophage phenotypes. Methods: NR8383 rat alveolar macrophages were stimulated with a mix of cytokines (LPS/IFN-γ or IL-4) for 24 h. The cells were further exposed to FAM inducing-compounds amiodarone and staurosporine. Following 24 h incubation, phagocytosis and lipid accumulation were measured using flow cytometry and high content image analysis techniques. The alveolar macrophages responses after exposure to cytokines were assessed by evaluation: (i) cell surface and biochemical markers such as: nitric oxide production, arginase-1 activity and MRC-1 receptor expression (ii) cellular morphology (iii) cellular functionality (phagocytic activity and lipids accumulation). Results: Macrophages activated with LPS/IFN-γ showed distinct morphological (increased vacuolation) features and functionality (increased lipidosis, decreased phagocytic activity). Foamy macrophage phenotypes induced by amiodarone also displayed characteristics of proinflammatory macrophages (significantly increased nitric oxide production, increased vacuolation and lipidosis and decreased phagocytosis). In contrast, staurosporine treatment resulted in increased NO production, as well as arginase-1 activity. Conclusion: High content image analysis was able to determine distinct differences in morphology between non-activated and LPS/IFN-γ activated macrophages, characterized by increased vacuolation and lipidosis. When exposed to compounds that induce a FAM phenotype, healthy non-activated macrophages displayed proinflammatory (amiodarone) or pro-apoptotic (staurosporine) characteristics but these responses were independent of a change in activation status. This technique could be applied in early drug discovery safety assessment to identify immune responses earlier and increase the understanding of alveolar macrophage responses to new molecules challenge in development of new inhalation therapies, which in turn will enhance decision-making in an early safety assessment of novel drug candidates.

Solid-phase microextraction for assessment of plasma protein binding, a complement to rapid equilibrium dialysis.

Aim: Determination of plasma protein binding (PPB) is considered vital for better understanding of pharmacokinetic and pharmacodynamic activities of drugs due to the role of free concentration in pharmacological response. Methodology & results: Solid-phase microextraction (SPME) was investigated for measurement of PPB from biological matrices and compared with a gold standard approach (rapid equilibrium dialysis [RED]). Discussion & conclusion: SPME-derived values of PPB correlated well with literature values, and those determined by RED. Respectively, average protein binding across three concentrations by RED and SPME was 33.1 and 31.7% for metoprolol, 89.0 and 86.6% for propranolol and 99.2 and 99.0% for diclofenac. This study generates some evidence for SPME as an alternative platform for the determination of PPB.

Water Uptake by Evaporating pMDI Aerosol Prior to Inhalation Affects Both Regional and Total Deposition in the Respiratory System.

As pulmonary drug deposition is a function of aerosol particle size distribution, it is critical that the dynamics of particle formation and maturation in pMDI sprays in the interim between generation and inhalation are fully understood. This paper presents an approach to measure the evaporative and condensational fluxes of volatile components and water from and to solution pMDI droplets following generation using a novel technique referred to as the Single Particle Electrodynamic Lung (SPEL). In doing so, evaporating aerosol droplets are shown capable of acting as condensation nuclei for water. Indeed, we show that the rapid vaporisation of volatile components from a volatile droplet is directly correlated to the volume of water taken up by condensation. Furthermore, a significant volume of water is shown to condense on droplets of a model pMDI formulation (hydrofluoroalkane (HFA), ethanol and glycerol) during evaporative droplet ageing, displaying a dramatic shift from a core composition of a volatile species to that of predominantly water (non-volatile glycerol remained in this case). This yields a droplet with a water activity of 0.98 at the instance of inhalation. The implications of these results on regional and total pulmonary drug deposition are explored using the International Commission of Radiological Protection (ICRP) deposition model, with an integrated semi-analytical treatment of hygroscopic growth. Through this, droplets with water activity of 0.98 upon inhalation are shown to produce markedly different dose deposition profiles to those with lower water activities at the point of inspiration.

The Influence of Oily Vehicle Composition and Vehicle-Membrane Interactions on the Diffusion of Model Permeants across Barrier Membranes.

In many instances, one or more components of a pharmaceutical or cosmetic formulation is an oil. The aims of this study were two-fold. First, to examine the potential of preferential uptake of one oily vehicle component over another into a model barrier membrane (silicone) from blended vehicles (comprising two from the common excipients isohexadecane (IHD), hexadecane (HD), isopropyl myristate (IPM), oleic acid (OA) and liquid paraffin). Second, to study the effect of membrane-vehicle interactions on the diffusion of model permeants (caffeine (CF), methyl paraben (MP) and butyl paraben (BP)) from blended vehicles. Selective sorption and partition of some oils (especially IHD and IPM) at the expense of other oils (such as OA) was demonstrated to take place. For example, the membrane composition of IHD was enriched compared to a donor solution of IHD-OA: 41%, 63% and 82% IHD, compared to donor solution composition of 25%, 50% and 75% IHD, respectively. Pre-soaking the membrane in IHD, HD or LP, rather than phosphate buffer, enhanced the flux of MP through the membrane by 2.6, 1.7 and 1.3 times, respectively. The preferential sorption of individual oil components from mixtures altered the barrier properties of silicone membrane, and enhanced the permeation of CF, MP and BP, which are typically co-formulated in topical products.

Multivariate Analytical Approaches to Identify Key Molecular Properties of Vehicles, Permeants and Membranes That Affect Permeation through Membranes.

There has been considerable recent interest in employing computer models to investigate the relationship between the structure of a molecule and its dermal penetration. Molecular permeation across the epidermis has previously been demonstrated to be determined by a number of physicochemical properties, for example, the lipophilicity, molecular weight and hydrogen bonding ability of the permeant. However little attention has been paid to modeling the combined effects of permeant properties in tandem with the properties of vehicles used to deliver those permeants or to whether data obtained using synthetic membranes can be correlated with those obtained using human epidermis. This work uses Principal Components Analysis (PCA) to demonstrate that, for studies of the diffusion of three model permeants (caffeine, methyl paraben and butyl paraben) through synthetic membranes, it is the properties of the oily vehicle in which they are applied that dominated the rates of permeation and flux. Simple robust and predictive descriptor-based quantitative structure-permeability relationship (QSPR) models have been developed to support these findings by utilizing physicochemical descriptors of the oily vehicles to quantify the differences in flux and permeation of the model compounds. Interestingly, PCA showed that, for the flux of co-applied model permeants through human epidermis, the permeation of the model permeants was better described by a balance between the physicochemical properties of the vehicle and the permeant rather than being dominated solely by the vehicle properties as in the case of synthetic model membranes. The important influence of permeant solubility in the vehicle along with the solvent uptake on overall permeant diffusion into the membrane was substantiated. These results confirm that care must be taken in interpreting permeation data when synthetic membranes are employed as surrogates for human epidermis; they also demonstrate the importance of considering not only the permeant properties but also those of both vehicle and membrane when arriving at any conclusions relating to permeation data.

Investigating the Suitability of High Content Image Analysis as a Tool to Assess the Reversibility of Foamy Alveolar Macrophage Phenotypes In Vitro.

Many potential inhaled medicines fail during development due to the induction of a highly vacuolated or "foamy" alveolar macrophage phenotype response in pre-clinical studies. There is limited understanding if this response to an inhaled stimulus is adverse or adaptive, and additionally if it is a transient or irreversible process. The aim of this study was to evaluate whether high content image analysis could distinguish between different drug-induced foamy macrophage phenotypes and to determine the extent of the reversibility of the foamy phenotypes by assessing morphological changes over time. Alveolar-like macrophages derived from the human monocyte cell line U937 were exposed for 24 h to compounds known to induce a foamy macrophage phenotype (amiodarone, staurosporine) and control compounds that are not known to cause a foamy macrophage phenotype in vitro (fluticasone and salbutamol). Following drug stimulation, the cells were rested in drug-free media for the subsequent 24 or 48 h. Cell morphometric parameters (cellular and nuclear area, vacuoles numbers and size) and phospholipid content were determined using high content image analysis. The foamy macrophage recovery was dependent on the mechanism of action of the inducer compound. Amiodarone toxicity was associated with phospholipid accumulation and morphometric changes were reversed when the stimulus was removed from culture environment. Conversely cells were unable to recover from exposure to staurosporine which initiates the apoptosis pathway. This study shows that high content analysis can discriminate between different phenotypes of foamy macrophages and may contribute to better decision making in the process of new drug development.

Wurster Fluidised Bed Coating of Microparticles: Towards Scalable Production of Oral Sustained-Release Liquid Medicines for Patients with Swallowing Difficulties.

Suspension of microparticles in an easy-to-swallow liquid is one approach to develop sustained-release formulations for children and patients with swallowing difficulties. However, to date production of sustained-release microparticles at the industrial scale has proven to be challenging. The aim of this investigation was to develop an innovative concept in coating sustained-release microparticles using industrial scalable Wurster fluidised bed to produce oral liquid suspensions. Microcrystalline cellulose cores (particle size <150 µm) were coated with Eudragit® NM 30 D and Eudragit® RS/RL 30 D aqueous dispersions using a fluidised bed coater. A novel approach of periodic addition of a small quantity (0.1% w/w) of dry powder glidant, magnesium stearate, to the coating chamber via an external port was applied throughout the coating process. This method significantly increased coating production yield from less than 50% to up to 99% compared to conventional coating process without the dry powder glidant. Powder rheology tests showed that dry powder glidants increased the tapped density and decreased the cohesive index of coated microparticles. Reproducible microencapsulation of a highly water-soluble drug, metoprolol succinate, was achieved, yielding coated microparticles less than 200 µm in size with 20-h sustained drug release, suitable for use in liquid suspensions. The robust, scalable technology presented in this study offers an important solution to the long-standing challenges of formulating sustained-release dosage forms suitable for children and older people with swallowing difficulties.

Comparison of Oral, Intranasal and Aerosol Administration of Amiodarone in Rats as a Model of Pulmonary Phospholipidosis.

'Foamy' alveolar macrophages (FAM) observed in nonclinical toxicology studies during inhaled drug development may indicate drug-induced phospholipidosis, but can also derive from adaptive non-adverse mechanisms. Orally administered amiodarone is currently used as a model of pulmonary phospholipidosis and it was hypothesized that aerosol administration would produce phospholipidosis-induced FAM that could be characterized and used in comparative inhalation toxicology. Han-Wistar rats were given amiodarone via (1) intranasal administration (6.25 mg/kg) on two days, (2) aerosol administration (3 mg/kg) on two days, (3) aerosol administration (10 mg/kg) followed by three days of 30 mg/kg or (4) oral administration (100 mg/kg) for 7 days. Alveolar macrophages in bronchoalveolar lavage were evaluated by differential cell counting and high content fluorescence imaging. Histopathology and mass-spectrometry imaging (MSI) were performed on lung slices. The higher dose aerosolised amiodarone caused transient pulmonary inflammation (p < 0.05), but only oral amiodarone resulted in FAM (p < 0.001). MSI of the lungs of orally treated rats revealed a homogenous distribution of amiodarone and a putative phospholipidosis marker, di-22:6 bis-monoacylglycerol, throughout lung tissue whereas aerosol administration resulted in localization of both compounds around the airway lumen. Thus, unlike oral administration, aerosolised amiodarone failed to produce the expected FAM responses.

Computational modelling and experimental validation of drug entrainment in a dry powder inhaler.

In a passive dry powder inhaler (DPI) a patient inhales to entrain drug powder. The goal of this study is to demonstrate experimentally that an Eulerian-Eulerian (EE CFD) computational fluid dynamics (CFD) method can accurately predict the entrainment of the dry powder formulation in DPIs. A CFD method that makes accurate predictions of the entrainment process can be applied in DPI design and optimization processes. Three different DPI entrainment geometries were tested. For each geometry, a transparent entrainment module was prepared. In each experiment, the chosen entrainment module was first filled with lactose powder and attached to an inhalation simulator (a computer controlled pump). The entrainment process was recorded with a high-speed camera. The resulting video footage was analysed and compared with CFD predictions. The observed distribution of powder in the entrainment compartment and the measured rate of drug entrainment were in good agreement with CFD predictions. Through a process of experimental validation, this study established the first demonstration that two-dimensional EE CFD methodology provides robust and accurate predictions of aerosol generation from DPI entrainment chambers. The findings support the wider application of EE CFD for the design optimization of DPI devices.

Pulmonary aerosol delivery and the importance of growth dynamics.

Aerosols are dynamic systems, responding to variations in the surrounding environmental conditions by changing in size, composition and phase. Although, widely used in inhalation therapies, details of the processes occurring on aerosol generation and during inhalation have received little attention. Instead, research has focused on improvements to the formulation of the drug prior to aerosolization and the resulting clinical efficacy of the treatment. Here, we highlight the processes that occur during aerosol generation and inhalation, affecting aerosol disposition when deposited and, potentially, impacting total and regional doses. In particular, we examine the response of aerosol particles to the humid environment of the respiratory tract, considering both the capacity of particles to grow by absorbing moisture and the timescale for condensation to occur. [Formula: see text].

A personalized medicine approach to the design of dry powder inhalers: Selecting the optimal amount of bypass.

In dry powder inhalers (DPIs) the patient's inhalation manoeuvre strongly influences the release of drug. Drug release from a DPI may also be influenced by the size of any air bypass incorporated in the device. If the amount of bypass is high less air flows through the entrainment geometry and the release rate is lower. In this study we propose to reduce the intra- and inter-patient variations of drug release by controlling the amount of air bypass in a DPI. A fast computational method is proposed that can predict how much bypass is needed for a specified drug delivery rate for a particular patient. This method uses a meta-model which was constructed using multiphase computational fluid dynamic (CFD) simulations. The meta-model is applied in an optimization framework to predict the required amount of bypass needed for drug delivery that is similar to a desired target release behaviour. The meta-model was successfully validated by comparing its predictions to results from additional CFD simulations. The optimization framework has been applied to identify the optimal amount of bypass needed for fictitious sample inhalation manoeuvres in order to deliver a target powder release profile for two patients.

Dispersing the Mists: An Experimental History of Medicine Study into the Quality of Volatile Inhalations.

BACKGROUND: Dr. Nelson's Improved Inhaler was first marketed with an advertisement in The Lancet in 1865. Revolutionary at the time for its ease of use and patient-friendliness, the inhaler is still in use for self-treatment by many all over the world. On the occasion of its 150th anniversary, this study reports an experimental historical medicine approach to identify evidence for the quality of vapor inhalers. METHODS: Through accessing reviews of the device's use by the contemporary medical establishment, it was established that Dr. Nelson's Inhaler enjoyed a reputation of quality and efficacy among reputable physicians generating empirical evidence of clinical performance. There was a general absence of product performance tests during this period. Therefore, modern inhalation performance testing was applied to test the aerosol delivery performance for Friars' Balsam, and its key chemical constituent, benzoic acid (BA). RESULTS: A respirable dose of 59.9 ± 9.0 µg of BA was aerosolized in a 10 minutes period from a dose of 3.3 mL Friars' Balsam (equivalent to 35.1 ± 0.2 mg of BA) in 375 mL of steaming water using the glass twin stage impinger at a flow rate of 60 L·min-1. The respirable dose from a standardized aqueous BA inhalation formulation increased from 115.9 ± 10.6 to 200.2 ± 19.9 µg by increasing the simulated inhalation period from 5 to 10 minutes. When tested with a simulated inhalation maneuver (500 mL tidal volume, 13 minutes-1 respiration rate, 1:2 inspiratory:expiratory ratio) a respirable dose of 112.8 ± 40.3 µg was produced. CONCLUSIONS: This work has highlighted the potential for aerosol drug delivery using steam inhalers that are popular with patients. Physicians should therefore be aware of the potential for lung dosing with irritants when patients self-medicate using the Nelson Inhaler with vaporizing formulations such as Friars' Balsam.

Predicting the Fine Particle Fraction of Dry Powder Inhalers Using Artificial Neural Networks.

Dry powder inhalers are increasingly popular for delivering drugs to the lungs for the treatment of respiratory diseases, but are complex products with multivariate performance determinants. Heuristic product development guided by in vitro aerosol performance testing is a costly and time-consuming process. This study investigated the feasibility of using artificial neural networks (ANNs) to predict fine particle fraction (FPF) based on formulation device variables. Thirty-one ANN architectures were evaluated for their ability to predict experimentally determined FPF for a self-consistent dataset containing salmeterol xinafoate and salbutamol sulfate dry powder inhalers (237 experimental observations). Principal component analysis was used to identify inputs that significantly affected FPF. Orthogonal arrays (OAs) were used to design ANN architectures, optimized using the Taguchi method. The primary OA ANN r2 values ranged between 0.46 and 0.90 and the secondary OA increased the r2 values (0.53-0.93). The optimum ANN (9-4-1 architecture, average r2 0.92 ± 0.02) included active pharmaceutical ingredient, formulation, and device inputs identified by principal component analysis, which reflected the recognized importance and interdependency of these factors for orally inhaled product performance. The Taguchi method was effective at identifying successful architecture with the potential for development as a useful generic inhaler ANN model, although this would require much larger datasets and more variable inputs.