ENOS 2015.

Methods	Single‐blind, parallel‐group, partial factorial study Randomisation via password‐protected, data‐encrypted website (minimisation by age, sex, stroke severity, time to treatment and total anterior circulation syndrome)
Participants	International (23 countries), multicentre (173 sites) 4011 participants: treatment 2000; control 2011 Age: treatment 70 years; control 70 years Male: treatment 1147 (57%); control 1150 (57%) Inclusion: IS or ICH; motor deficit in arm or leg, or both; systolic BP 140 to 220 mmHg Time to randomisation: within 48 hours of onset
Interventions	Treatment: transdermal GTN patch 5 mg once daily Control: no GTN (blinding with gauze dressing over patch/equivalent area of skin) In appropriate participants: Treatment: continue pre‐stroke antihypertensive medication Control: stop pre‐stroke antihypertensive medication Duration: 7 days
Outcomes	Primary: mRS at day 90 Secondary: days 1 to 7: BP and HR day 7: recurrent stroke discharge: length of hospital stay; disposition day 90: death or dependency (mRS > 2); BI; EQ‐5D, EQ‐VAS; MMSE, TICS‐M; animal naming; Zung Depression Scale
Notes	Exclusions: GCS < 8; pure sensory stroke; preceding dependency (mRS 3 to 5); confounding neurological or psychiatric illness; stroke mimic; severe liver or renal dysfunction; severe comorbidity; pregnant or breastfeeding; planned surgical intervention; previous participation in ENOS 2015; contraindication to or definite need for nitrates and/or prestroke antihypertensive medication
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Central computer randomisation with minimisation
Allocation concealment (selection bias)	Low risk	Central website‐based
Blinding (performance bias and detection bias) All outcomes	Low risk	GTN was given single‐blinded (participant), whilst outcomes were assessed centrally and blinded to treatment group
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	GTN was given single‐blinded as no placebo patches were available. A gauze dressing was placed over the GTN patch or equivalent area of skin out of sight. As a result, participants were blinded whilst the treating clinician was unblinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcomes assessed centrally, blinded to treatment group
Incomplete outcome data (attrition bias) All outcomes	Low risk	No difference between trial groups
Selective reporting (reporting bias)	Low risk	All prespecified outcomes reported
Other bias	Low risk	None found