Wildiers 2009.

Methods	Open‐label randomised phase III randomised multi‐centre trial
Participants	440 patients randomised: 333 enrolled; remainder excluded because no informed consent or not met inclusion criteria. 333 patients allocated to Group I (115 patients); Group 2 (112 patients) or Group 3 (106 patients). Mean age: Group 1 = 70.7 years Group 2 = 74.3 years Group 3 = 72.6 years Gender: approximately 50:50 in all three groups All had cancer except 5 in Group 1, 7 in Group 2 and 5 in Group 3.
Interventions	Randomly allocated to: Group 1: Atropine 0.5 mg subcutaneous bolus, followed by 3 mg/24 hours Group 2: Scopolamine (hyoscine hydrobromide) 0.25 mg subcutaneous bolus, followed by 1.5 mg/24 hours Group 3: Hyoscine butylbromide 20 mg subcutaneous bolus, followed by 60 mg/24 hours If death rattle persisted at score of 2 or 3 after 12 hours, starting bolus dose of same drug readministered and maintenance dose doubled.
Outcomes	Nurse assessment at 30 minutes, 1 hour, 4 hours, 12 hours, 24 hours and every 24 hours until death. Death rattle intensity scored using scale of 0‐3: 0 = not audible; 1 = only audible near the patient; 2 = clearly audible at the end of the patient's bed in a quiet room; 3 = clearly audible at a distance of about 9.5m in a quiet room. At one hour: no significant difference in effectiveness between the three groups: Group 1 (atropine): 42% Group 2 (scopolamine): 37% Group 3 (hyoscine butylbromide): 42%. Steady increase in effectiveness up to 24 hours, from 70% of patients at start of therapy to 30% at 24 hours had death rattle intensity scores of 2 or 3.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Stratified per centre
Allocation concealment (selection bias)	Low risk	Closed envelope system
Selective reporting (reporting bias)	Low risk	Outcomes fully reported.
Incomplete outcome data addressed	Unclear risk	Randomisation took place ahead of consent and checking against inclusion criteria ‐ analysis not carried out on basis of intention to treat
Blinding	High risk	Open label