(M1230-08-50) Predicting Lung Dose and Asthma Control from Real-World Dry Powder Inhaler Use

Purpose: Asthma remains a major global health burden, with most patients relying on daily dry powder inhalers (DPIs) for symptom control. DPIs are breath-actuated devices whose effectiveness depends on patient-specific inhalation parameters.^{1, 2} These parameters govern powder dispersion, pulmonary drug deposition, and ultimately influence clinical outcomes, as reflected in Asthma Control Test (ACT) scores. Despite their widespread use, most in vitro assessments rely on standard flow profiles and do not reflect real-world patient inhalation behavior³. As a result, there is limited evidence linking in vitro drug delivery performance to clinical efficacy. To address this gap, this study aimed to develop an in vitro predictive model using real-world inhalation profiles to estimate dose delivery from DPI and to evaluate its association with asthma control outcomes. By integrating acoustic flow monitoring of at-home DPI use with in vitro aerodynamic testing, we sought to provide a physiologically relevant tool for predicting DPI performance under actual use conditions.

Methods: In a 12-week prospective study, inhalation flow profiles were collected from 107 adult asthma patients during routine twice-daily use of the Symbicort Turbuhaler^®. An acoustic device attached on the DPI was used at home to record inhalation profiles, from which key flow parameters, including peak inspiratory flow (PIF), inhaled volume, and rise time to PIF, were extracted for each actuation. To estimate the lung dose associated with these real-world profiles, a surface response model was established using a three-factor, three-level Box-Behnken design (BBD). This design evaluated the effects of PIF (20-100 L/min), inhaled volume (0.5-2 L), and rise time to PIF (0-3 sec) on in vitro budesonide deposition. In vitro experiments were conducted to quantify emitted dose and lung dose (160 µg/actuation) from the Symbicort Turbuhaler^® across 15 simulated profiles, using an Alberta Idealized Throat connected to a mixing inlet, pre-separator, Next Generation Impactor (NGI), and breathing simulator ⁴ (Figure 1). The emitted dose (%), defined as the dose released from the inhaler to the labeled dose, and the lung dose (%), calculated as the fraction of the emitted dose captured by the NGI , were determined. The resulting dataset was used to generate a predictive model linking real-world inhalation parameters to budesonide emitted dose (BUDED) and lung dose (BUDLD). Each patient's profile was then input into the model to estimate individual BUDLD, which was subsequently correlated with changes in ACT scores to assess clinical relevance.

Results: BBD analysis identified PIF ( p < 0.001) and inhaled volume (p = 0.0254) as the primary determinants of BUDED, with rise time to PIF showing a marginal effect (p = 0.0692). For BUDLD, PIF (p < 0.001) and rise time (p = 0.015) were the dominant factors, while inhaled volume had a lesser impact (p = 0.077). These findings suggest that PIF strongly influenced both powder emission and lung deposition, whereas rise time primarily affected lung deposition by modulating airflow acceleration. Specifically, higher PIF and larger inhaled volume could facilitate powder fluidization and aerosolization, facilitating greater drug emission from the inhaler. Additionally, higher PIF and shorter rise times could increase airflow turbulence and impaction force within the inhaler, improving particle deagglomeration and generating a greater fraction of smaller particles capable of penetrating through the upper airway. The final regression model demonstrated strong predictive performance (R² = 0.89), indicating that profiles with PIF > 60 L/min and rise time < 1 sec were associated with significantly greater BUDLD (Figure 2). Among 107 patients, predicted monthly BUDLD was positively correlated with ACT score improvements (r = 0.62, p < 0.01). Patients who consistently achieved PIF > 60 L/min experienced a mean ACT score increase of 4.2 points over the 12-week period. Notably, individuals with a predicted BUDLD exceeding 3000 µg per month were significantly less likely to experience ACT score deterioration (odds ratio 0.36, 95% CI 0.13–0.96, p = 0.042; Figure 3), supporting the clinical relevance of the model.

Conclusion: This study presents the first in vitro predictive model that integrates real-world inhalation profiles with estimations of both emitted dose and lung dose, alongside clinical outcomes in asthma. The model provides a physiologically relevant assessment of drug delivery, reflecting patient-specific inhalation behavior, and exhibits robust predictive accuracy under actual use conditions. Notably, predicted BUDLD was strongly correlated with ACT score improvements, while BUDED highlighted the critical role of PIF and inhaled volume in effective powder dispersion and lung deposition. Additionally, controlling rise time to achieve optimal inspiratory airflow dynamics was identified as critical for maximizing lung dose. These findings underscore the potential of this approach to optimize DPI delivery and guide patient training strategies. Future research should validate this methodology across diverse inhaler and patient populations to further advance personalized DPI-based therapy.

References: (1) W E S Kamin, T. G., S Roeder, G Scheuch, T Trammer, R Juenemann, R M Cloes. Mass Output and Particle Size Distribution of Glucocorticosteroids Emitted from Different Inhalation Devices Depending on Various Inspiratory Parameters. J Aerosol Med,. 2002, 15, 65–73. DOI: 10.1089/08942680252908593.
(2) Azouz, W.; Chetcuti, P.; Hosker, H. S.; Saralaya, D.; Stephenson, J.; Chrystyn, H. The inhalation characteristics of patients when they use different dry powder inhalers. J Aerosol Med Pulm Drug Deliv 2015, 28 (1), 35-42. DOI: 10.1089/jamp.2013.1119 From NLM Medline.
(3) Bagherisadeghi, G.; Larhrib, E. H.; Chrystyn, H. Real life dose emission characterization using COPD patient inhalation profiles when they inhaled using a fixed dose combination (FDC) of the medium strength Symbicort((R)) Turbuhaler((R)). Int J Pharm 2017, 522 (1-2), 137-146. DOI: 10.1016/j.ijpharm.2017.02.057 From NLM Medline.
(4) Olsson, B.; Borgstrom, L.; Lundback, H.; Svensson, M. Validation of a general in vitro approach for prediction of total lung deposition in healthy adults for pharmaceutical inhalation products. J Aerosol Med Pulm Drug Deliv 2013, 26 (6), 355-369. DOI: 10.1089/jamp.2012.0986 From NLM Medline.

Acknowledgements: Informed consent has been obtained from all patients included in this study. To the best of our knowledge, the named authors have no conflict of interest, financial or otherwise.

Figure 1. Inhalation profile simulation system setup

Figure 2. Three-dimensional surface plots illustrating the effects of PIF (20–100 L/min), inhaled volume (0.5–2 L), and rise time to PIF (0, 1.5, and 3 s) on BUDED and BUDLD from the Symbicort Turbuhaler®. Volumeinh: inhaled volume.

Figure 3. Proportion of ACT deterioration stratified by 4-week BUDLD dose. Bar chart showing the percentage of patients with ACT Change < −3, grouped by a 4-week BUDLD threshold of 3,000 μg. The proportion of ACT deterioration was significantly higher in the low-dose group (≤3,000 μg) compared to the high-dose group (>3,000 μg) (p = .0238, Fisher’s exact test). Sample sizes are indicated for each group.