Abstract Rationale The bronchodilator responsiveness (BDR) testing is widely used to differentiate asthma from COPD and often serves as an inclusion criterion in clinical trials. However, its validity is questioned1. This study assesses the discriminative performance of BDR and compares it with spirometry interpreted by ArtiQ.Spiro2, an AI-based support software predicting among five diagnoses based on demographics, smoking history and spirometry data. Methods A retrospective analysis was performed on adults who underwent spirometry with BDR testing at the University Hospital of Leuven between 2014 and 2018. From this cohort, 287 patients were selected based on a clinical diagnosis of either asthma (n = 110) or COPD (n = 177), established through comprehensive pulmonary evaluations. The proportion of patients with a positive BDR was calculated according to the 2005 ATS/ERS criteria (ΔFEV1 or ΔFVC12% and 200 mL), the 2021 ATS/ERS criteria (ΔFEV1 or ΔFVC10% predicted), and the higher-confidence definition of GINA (ΔFEV1 or ΔFVC≥15% and ≥400 mL). Analyses were stratified by lung pattern per the 2021 ATS/ERS interpretation standard, with differences assessed using Fisher’s exact test. The performance of ArtiQ.Spiro and the potential added value of BDR testing were also evaluated. Results According to the 2005 ATS/ERS criteria, 20% of asthma and 22% of COPD patients showed a positive BDR. These proportions declined to 16% and 18% with the 2021 ATS/ERS criteria, and to 7% and 6% under the GINA criteria. Among COPD patients, 90% showed an obstructive pre-bronchodilator spirometry lung pattern. Asthma cases were mainly obstructive (37%) or normal (54%). When comparing obstructive subgroups, the 2005 ATS/ERS BDR was non-significantly higher in asthma than COPD (32% vs. 24%, p = 0.319) and significantly higher in obstructive than normal asthma patterns (32% vs. 14%, p = 0.045). ArtiQ.Spiro achieved 76% overall accuracy on this dataset. BDR was found in 11% of misclassified asthma cases and 25% of correctly classified COPD cases, indicating limited potential added value to the model. Conclusions Although the BDR test is interpreted alongside clinical information, its prevalence is low and similar in asthma and COPD. ArtiQ.Spiro, using pre-bronchodilator spirometry, demonstrated good discriminative performance, while BDR provided no direct additional value. Therefore, care should be taken when relying on BDR as an eligibility criterion in clinical trials, as it may inappropriately restrict patient inclusion. Moreover, the increasing recognition asthma and COPD overlap further supports the need for a more integrated approach to patient selection. References: 1Beasley, AJRCCM, 2023 https://doi.org/10.1164/rccm.202308-1436OC 2Sunjaya, ERJ Open Res., 2025 https://doi.org/10.1183/23120541.00116-2025 This abstract is funded by: Clario
Rakic et al. (Fri,) studied this question.