| First author [ref.] | Year | Country | Study design | Interface | Received NIV | NIV failure | Transmission among HCW# | Mortality | Observations |
| Lin [11] | 2003 | China | Retrospective, single centre | Face mask | n=40 (51.9%) | n=8 (10.3%) | No | 9% | 70 (90.9%) patients were clinically cured |
| Cheung [12] | 2004 | China | Case series, single centre (n=31) | Face mask | n=20 | n=14 (70%) | No | 0% | |
| Fowler [13] | 2004 | Canada | Retrospective, single centre | Face mask | 0 | n=38 | No | 50% after NIV failure | Affected patients had primarily single organ respiratory failure |
| Sung [14] | 2004 | China | Prospective, single centre (n=37) | Face mask | n=15 | n=21 (15.2%) | No | n=15 (10.9%) | Most patients had significant comorbidities |
| Yam [8] | 2005 | China | Retrospective, single centre | Face mask | n=21 | n=8 (38%) | No | n=9 (35%) | Early application of NIV as initial support for SARS-related ARF appeared to be associated with significantly reduced need for ETI and mortality |
HCW: healthcare workers; ARF: acute respiratory failure; ETI: endotracheal intubation. #: instances of transmission of SARS among HCW.