Schmied 1996.

Methods	Design: RCT Operative phase: intra + post‐operative Withdrawals: none Setting: multicentre (2) (Austria and USA) Sample size: 60 Funding: Augustine Medical Inc, the Joseph Drown and Max Kade Foundations, and National Institutes of Health grant RO 1GM49670. Mallinckrodt Anesthesia Products Inc, donated thermocouples and thermometers. The authors do not consult for, accept honoraria from, or own shares or share options in any anaesthesia‐ or surgery‐related company
Participants	Age (mean): 63 years Gender (M/F): 23/37 ASA grade: I ‐ III Surgery type: elective (primary unilateral total hip arthroplasty) Surgery duration (mean range): < 2 hrs (85/87 mins) Anaesthesia type: general
Interventions	Intervention (ABSW): n = 30 Upper‐body forced‐air warming cover (Bair‐Hugger®, Augustine Medical, Eden Prairie, MN) and a warmer set to "high" + intravenous fluids warmed to 37ºC Duration (mean): not stated Temperature: 37ºC Body area covered: upper body Proportion covered: not stated Control: n = 30 "Hypothermia group" (active skin and fluid warming was avoided) Co‐interventions: Low‐molecular weight heparin (5000 IU every 8 hrs) starting 2 hrs before surgery Room temperature: 21ºC
Outcomes	Fluids infused (crystalloids; colloids) (intraoperatively and postoperatively) Blood loss (cumulative) (ml) N of participants transfused Transfusion (RBC) (ml/participant) Other outcomes reported not included in the review: Haemoglobin (end of surgery; next morning) Hematocrit Prothrombin and plasma thrombin time Arterial blood pressure Heart rate Blood platelet number
Notes	Comparison 1
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization was based on computer‐generated codes sealed in sequentially‐numbered, opaque envelopes
Allocation concealment (selection bias)	Low risk	Randomization was based on computer‐generated codes sealed in sequentially‐numbered, opaque envelopes
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not reported
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Baseline comparability of groups	Low risk	To a high extent according to Table 1
Co‐interventions equal between groups	Low risk	Yes
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomized participants were analysed
Selective reporting (reporting bias)	Unclear risk	We did not have access to the protocol, therefore we cannot exclude risk of selective reporting with the information provided
Other bias	Low risk