Objectives We compared the risk of environmental contamination among patients with COVID-19 who received high-flow nasal cannula (HFNC), noninvasive ventilation (NIV), and conventional oxygen therapy (COT) via nasal cannula for respiratory failure.Methods Air was sampled from the hospital isolation rooms with 12 air changes/hr where 26 patients with COVID-19 received HFNC (up to 60 l/min, n = 6), NIV (n = 6), or COT (up to 5 l/min of oxygen, n = 14). Surface samples were collected from 16 patients during air sampling.Results Viral RNA was detected at comparable frequency in air samples collected from patients receiving HFNC (3/54, 5.6%), NIV (1/54, 1.9%), and COT (4/117, 3.4%) ( P  = 0.579). Similarly, the risk of surface contamination was comparable among patients receiving HFNC (3/46, 6.5%), NIV (14/72, 19.4%), and COT (8/59, 13.6%) ( P  = 0.143). An increment in the cyclic thresholds of the upper respiratory specimen prior to air sampling was associated with a reduced SARS-CoV-2 detection risk in air (odds ratio 0.83 [95% confidence interval 0.69-0.96],  P  = 0.027) by univariate logistic regression.Conclusion No increased risk of environmental contamination in the isolation rooms was observed in the use of HFNC and NIV vs COT among patients with COVID-19 with respiratory failure. Higher viral load in the respiratory samples was associated with positive air samples.