Avdeev 1998.

Methods	Country: Russia Design: Randomised controlled parallel trial. Participants were matched for demographic and physiological norm values Study site: One hospital in Moscow, conducted between September 1995 and March 1997 Method of analysis: Unclear Aim: To determine the effect of NIV on need for endotracheal intubation, mortality rate, length of hospital stay, and incidence of complications in patients with acute respiratory failure caused by AECOPD
Participants	Eligible for study: Not stated Recruited: 58 adult patients with acute respiratory failure due to AECOPD (29 in each group) Completed: Not stated Age: NIV group: mean (SD) age = 63.4 (5.5) years; usual care group: mean (SD) age = 66.2 (7.1) years Gender: NIV group (M:F) = 26:3, usual care group (M:F) = 22:7 Criteria used to define COPD: Not stated Inclusion criteria: Insufficient information available Exclusion criteria: Insufficient information available
Interventions	Intervention description: NIV plus usual care. BiPAP ventilators (Respironics, Inc., Murrysville, PA, USA) used with inspiratory pressure titrated to 30 cmH2O and expiratory pressure of 4 to 6 cmH2O. Both face masks and nasal masks were used Control description: Oxygen, bronchodilators, steroids, and theophylline Duration of intervention: Mean (SD) duration of NIV was 29 (25) hours Intervention delivery by: Insufficient information available Setting: Intermediate care unit
Outcomes	Method of outcome data collection: Not clear Prespecified primary outcomes: Not clear which outcomes deemed primary. Specified outcomes were need for intubation, mortality rate, length of hospital stay, and incidence of complications. No clinical trial registry to confirm Prespecified secondary outcomes: Other outcomes reported included breathlessness score (Borg) and ABGs Follow‐up period: Data collected until hospital discharge
Notes	Trial was published in Russian Study author was contacted and provided additional information Funder: Unclear
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Described as randomised; other information not available
Allocation concealment (selection bias)	Low risk	Sealed opaque envelopes were used for treatment allocation.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Owing to nature of the intervention, blinding was not possible. No sham NIV used; however, unlikely to have affected primary outcomes. May have affected subjective ratings of dyspnoea
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Unsure whether investigators were involved in participant treatment
Incomplete outcome data (attrition bias) All outcomes	High risk	3 participants from the NIV group were excluded owing to NIV intolerance. Insufficient information available to determine reasons or potential impact
Selective reporting (reporting bias)	Low risk	No evidence of selective reporting
Imbalance of outcome measures at baseline All outcomes	Unclear risk	Insufficient information available
Comparability of intervention and control group characteristics at baseline	Unclear risk	Participants in the NIV group had lower age at baseline. Unclear whether statistically significant. Insufficient information available for other outcomes
Protection against contamination All outcomes	Unclear risk	Insufficient information available
Selective recruitment of participants	Low risk	N‐values and methods of recruitment were similar
Other bias	Low risk	No other sources of bias identified