The study was randomized and included 340 patients with low anesthetic risk (ASA I–II) who underwent diagnostic flexible bronchoscopy under deep sedation. Patients were divided into a group with a conventional low-flow nasal cannula and three groups with a high-flow nasal cannula (HFNC) with flow rates of 25, 45 and 65 l/min. The incidence of desaturation (SpO₂ < 90% > 10 s) was 57.14% in the conventional cannula group, while it was significantly lower in the HFNC groups: 16.87% (25 L/min), 12.05% (45 L/min) and 5.00% (65 L/min). All HFNC groups had higher trough SpO₂ during exercise and lower CO₂ retention compared to conventional cannula. All bronchoscopies took approximately 5 minutes and did not require a laryngeal mask or intubation. The cough score was low (median VAS 1–1.5), corresponding to a well-suppressed cough reflex under deep sedation. More than 97% of patients expressed willingness to undergo bronchoscopy again. The authors conclude that HFNC significantly reduces the risk of desaturation and partially alleviates CO₂ accumulation during bronchoscopy in deep sedation without the need for intubation.