Background and objective: Hypoxemia and respiratory failure (RF) in chronic obstructive pulmonary disease (COPD) are associated with exacerbations, comorbidities and increased mortality. However, the prevalence of hypoxemia and RF in stable COPD is unknown. We aimed at investigating the prevalence and determining predictive factors for chronic gas exchange abnormalities in COPD patients. Methods: A retrospective cohort study that enrolled clinically stable COPD patients referring to a pulmonary outpatient clinic. Anthropometrics, clinical characteristics, blood gas analysis and lung function were analyzed. Patients were grouped according to hypoxemia (PaO₂ <80 and ≥60 mmHg), type 1 (PaO₂ < 60 mmHg) or type II (PaO₂ < 60 and PaCO₂ > 45 mmHg) RF. A sensitivity analysis adopting an age-adjusted definition of hypoxemia was performed. Predictive factors for hypoxemia or RF were assessed with multifactorial analysis. Results: We analyzed data from 515 patients. Fixed-ratio hypoxemia, RF type 1 and type 2 were observed in 352 (68.3%), 27 (5.2%) and 43 (8.3%) patients, respectively. Risk of hypoxemia was associated with preserved alveolar volume, residual volume/total lung capacity, and lung diffusion capacity. Heart failure, ischemic heart disease, atrial fibrillation, and metabolic syndrome were predictive factors for RF. Patients with age-adjusted hypoxemia (n = 321 patients, 62.3%) showed no difference in terms of anthropometrics, lung function, and clinical characteristics as compared with fixed-threshold hypoxemia. Conclusions: Hypoxemia is frequent in stable COPD. Lung function parameters and comorbidities can support the identification of patients at risk of RF. Blood gas analysis should be always performed in patients with COPD to allow for personalized therapy and management.

Background: International guidelines have published detailed recommendations on the use of EBUS-TBNA for staging and diagnosing lung cancer for lymph nodes (LNs) >10 mm, but provide limited guidance for sampling smaller LNs. Aim: The study evaluated the technical feasibility and clinical yield of EBUS-TBNA in patients with suspected lung disease and LNs <10 mm. Material & Methods This was a retrospective, observational, single-center study that included consecutive patients that underwent EBUS-TBNA at Luigi Sacco University Hospital, Milan (Italy) as part of the diagnostic workup for lung lesions. Mediastinal LNs were categorized as ≤10 mm or >10 mm based on short-axis diameter on chest CT. Sample adequacy was assessed by means of Rapid On-Site Evaluation. The primary outcome was technical feasibility of sampling LNs ≤10 mm, while anatomopathological yield was the secondary outcome. Results: A total of 195 LNs were sampled in 130 patients, and 48 (36.9%) patients had 86 (44.1%) LNs with a short-axis diameter ≤10 mm. The adequacy rate was 79/86 (91.9%). A specific cytopathological diagnosis was achieved in 24 LNs ≤ 10 mm (30.4%), allowing to carry out or rule out a diagnosis in 23 patients (47.9%). Specific diagnoses included: adenocarcinoma (n=4), small cell lung cancer (n=1), metastatic disease (n=2), sarcoidosis (n=6), and negative N2/N3 staging in patients with peripheral lung masses (n=10). Conclusions: EBUS-TBNA demonstrated high sample adequacy and clinical yield also in LNs ≤10 mm. These findings support extending EBUS-TBNA sampling beyond size-based criteria, particularly when integrating ultrasound features and clinical context, which may potentially improve early disease detection and favor therapeutic management and surgical planning.