Liu 2005.

Methods	Country: China, Nanjing Design: Randomised controlled trial. Randomisation method was not reported Study site: Single centre in Nanjing, from December 2001 to December 2003 Setting: ICU Methods of analysis: SPSS 11.5 for analysis. P < 0.05 statistically significant Aim: To evaluate effects of early use of non‐invasive positive pressure ventilation on gas exchange, rate of endotracheal intubation, and in‐hospital mortality among patients with acute exacerbations of COPD
Participants	Eligible for study: Not mentioned Recruited: 36 patients with acute exacerbations of COPD with pH from 7.25 to 7.35 and PaCO2 >45 mmHg were enrolled. 18 participants were randomised to NIV group and 18 to standard therapy group. Baseline characteristics were similar in both groups Completed: 18 participants in NIV group and 18 in usual care group Age: NIV group: 70.8 ± 5.1 years. Usual care group: 68.4 ± 6.0 years Gender: NIV group: 15 male and 3 female. Usual care group: 14 male and 4 female Criteria used to define COPD: COPD as defined in 1997 Chinese Association of Respiratory Physician COPD plan Inclusion criteria: Acute exacerbation of COPD according to history and physical examination with radiological findings; pH 7.23 to 7.35 and PaCO2 > 45 mmHg Exclusion criteria: Refusal for NIV; GCS < 8; pneumothorax; respiratory arrest; arrhythmia; multi‐organ failure; severe abdominal distension, bowel perforation or bleeding, recent bowel surgery; face trauma; face irregularities
Interventions	Intervention description: NIV provided by BiPAP vision using full face mask, with initial PEEP 4 cmH2O and pressure support 2 to 4 cmH2O, titrated by 2 cmH2O until RR < 28 bpm and SpO2 > 90%. All participants also received bronchodilator, antibiotics, mucolytics, and supplementary oxygen to maintain SpO2 > 90% Control description: Bronchodilator, antibiotics, mucolytics, supplementary oxygen to maintain SpO2 > 90% Duration of intervention: At least 3 days. Initial NIV maintained over 2 hours with aim of at least 8 hours per day Intervention delivery by: Not reported in text
Outcomes	Method of outcome data collection: Data were collected at baseline and throughout hospital stay Prespecified primary outcome: Protocol not available. In text outcomes: Endotracheal intubation rate, in‐hospital mortality rate, ABG changes Prespecified secondary outcome: Protocol not available. In text outcomes: NIV complications; RR and HR Validation: ABG, HR, RR Follow‐up period: Throughout admission, until discharged or end point reached Number of follow‐up periods reported on during study: Not reported Indications for intubation: pH < 7.25, increasing PCO2, PO2 < 45 mmHg, GCS < 8, cardiopulmonary arrest, RR < 8/min, or RR > 40/min Complications: NIV was not tolerated in 1 participant. One had a face pressure ulcer, which resolved after NIV was stopped
Notes	Mean duration: 6.1 ± 1.9 days; 53.7 ± 26.6 hours; mean 8.8 ± 3.6 hours per day Maximum PS 14.3 ± 2.8 cmH2O; maximum PEEP 4.3 ± 0.8 cmH2O Paper in Chinese, with limited translation from translator. Attempted to contact study authors for more information, but no reply Funder: Great Topic Foundation of Health Bureau Jiangsu Province, China (No. H200102)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random number table used for randomisation sequence generation
Allocation concealment (selection bias)	Unclear risk	Insufficient information reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Owing to the nature of the intervention, blinding was not possible. No sham NIV was used. However, outcomes reported were objective outcomes and were unlikely to be affected
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not mentioned who delivered the intervention
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Not mentioned how missing variables, if any, were handled. Unsure whether all participants completed the study
Selective reporting (reporting bias)	Unclear risk	Prespecified outcomes not clearly explained
Imbalance of outcome measures at baseline All outcomes	Unclear risk	Insufficient information reported
Comparability of intervention and control group characteristics at baseline	Unclear risk	Insufficient information reported
Protection against contamination All outcomes	Unclear risk	Insufficient information reported
Selective recruitment of participants	Unclear risk	Insufficient information reported
Other bias	Unclear risk	Limited information available to assess whether free of other sources of bias