Study	Reason for exclusion
ACCORD 2010	Study designed to test the effect of BP lowering in addition to glycaemic control in people with diabetes. Participants were assigned to 2 BP treatment goals ‐ intensive (SBP < 120 mmHg) or standard (SBP < 140 mmHg). Various classes of antihypertensive drugs used but recommended start with combination of diuretic and ACE inhibitor or beta‐blocker. Beta‐blockers not first‐line or monotherapy.
ADaPT 2008	Observational study conducted in primary care compared ACE inhibitor‐based treatment (ramipril) with a treatment based on diuretics or beta‐blockers. Not randomised.
APSIS 2006	Study compared verapamil or metoprolol in people with stable angina pectoris. Not all participants had hypertension (27%). Mean baseline BP not given.
CAPP 1999	This study compared the effects of ACE inhibitors and conventional therapy (diuretics and beta‐blockers) on cardiovascular morbidity and mortality in people with hypertension. Findings were not reported separately for beta‐blockers.
CAPRICORN 2001	Trial evaluated the effects of carvedilol with placebo on survival in post‐MI participants with left ventricular dysfunction with or without symptomatic heart failure. All participants given ACE inhibitors for at least 48 hours before randomisation. Not all participants had hypertension (54%) and beta‐blockers not first‐line or monotherapy.
CARDHIAC 2008	Study examined effects of doxazosin GITS and atenolol on 3 measures of target organ damage in people with type 2 diabetes and hypertension. Participants received ACE inhibitors or ARB and diuretic initially before receiving doxazosin GITS and atenolol. Beta‐blockers not first‐line or monotherapy.
CHHIPS 2009	This RCT, which was conducted in 6 centres in the UK, evaluated the effects of active treatment with the ACE inhibitor, lisinopril, or beta‐blocker, labetalol, compared to placebo in people aged > 18 years with a clinical diagnosis of suspected stroke (with symptom onset < 36 hours) and hypertension (defined as SBP > 160 mmHg). After 2 weeks of treatment, study participants were routinely started on an ACE inhibitor with or without a diuretic irrespective of whether they had normal BP or hypertension, unless they were deemed to be unsuitable for such therapy. Decisions with regard to future antihypertensive therapy were delayed until the end of the trial intervention (2 weeks). The proportion of participants on assigned treatment at the end of the study was 71% in the beta‐blocker group, 68% in the ACE inhibitor group, and 80% in the placebo group. 172 participants, with mean age 74 years, were enrolled and the study reported mortality data at 3 months. We excluded this study because of the short duration (i.e. only 2 weeks) of relevant interventions.
CIBIS‐II 1999	Trial compared bisoprolol and placebo in people with heart failure receiving standard therapy with an ACE inhibitor and diuretic. Not all participants had hypertension (mean baseline BP 139/80 mmHg) and beta‐blocker not first‐line or monotherapy.
COMET 2003	Trial compared carvedilol and metoprolol in people with chronic heart failure. Not all had hypertension (36%). Mean baseline BP 126/77 mmHg.
CONVINCE 1998	The Controlled ONset Verapamil Investigation of Cardiovascular Endpoints (CONVINCE) Trial is a randomised, prospective, double‐blind, parallel‐group, 2‐arm, multicentre, international trial. The study recruited 15,000 people with hypertension, aged > 55 years, with an established second risk factor for cardiovascular disease and followed them for 5 years to compare the effects of controlled onset‐extended release verapamil 180 mg/day and hydrochlorothiazide 12.5 mg/day or atenolol 50 mg/day. Data has not been reported separately for hydrochlorothiazide and atenolol.
COPE 2005	Study compared a combination of ARB, beta‐blocker, or thiazide diuretic in addition to a calcium‐channel blocker, benidipine hydrochloride, in Japanese people with hypertension. Beta‐blockers not first‐line treatment or monotherapy.
COPERNICUS 2004	Study compared carvedilol vs placebo in people with chronic heart failure and receiving spironolactone or not at baseline. Not all participants had hypertension (mean baseline BP 123/76 mmHg).
COSMOS 2010	People with stage 1 or 2 hypertension were randomised evenly to 1 of 15 groups for 6 weeks: extended‐release carvedilol (carvedilol CR) monotherapy 20 mg/day, 40 mg/day, or 80 mg/day; lisinopril monotherapy 10 mg/day, 20 mg/day, or 40 mg/day; or 1 of 9 combinations of carvedilol CR + lisinopril initiated simultaneously. The study has not reported effects on mortality or cardiovascular endpoints.
Dietz 2008	This RCT was conducted in 85 centres in China, Germany, India, South Africa, Spain, and Turkey. People with hypertension (defined as mean sitting DBP 95 mmHg to 110 mmHg) were randomised to once‐daily aliskiren 150 mg (231 participants), atenolol 50 mg (231 participants), or the combination (150/50 mg; 232 participants) for 6 weeks, followed by a further 6 weeks on double the initial doses of aliskiren and atenolol. Aliskiren is the first direct renin inhibitor to be approved for the treatment of hypertension. The proportion of participants on assigned treatment at the end of the study was 92.2% in the beta‐blocker group, 91.3% in the direct renin inhibitor group, and 88.4% in the combination group. The trial followed up 694 participants (mean age 55.2 years, 23% aged ≥ 65 years) for 12 weeks. We excluded this study because of the short duration (i.e. only 12 weeks) of relevant interventions.
Dutch TIA 1993	The trial evaluated the effects of a beta‐blocker (atenolol) in people after a transient ischaemic attack or non‐disabling ischaemic stroke in 56 collaborative centres in the Netherlands. Participants were randomised to atenolol or a matching placebo. The proportion of participants on assigned treatment in the beta‐blocker group was 71% at 2 years (and 64% at 3 years) and in the placebo group was 75% at 2 years (and 68% at 3 years). The trial followed up 1473 participants (52% aged > 65 years) for a mean duration of 2.7 years. We excluded the trial because only 29% of participants had hypertension at baseline.
GEMINI 2008	Trial compared effects of carvedilol with metoprolol on glycaemic control in people with hypertension and type‐2 diabetes. BP was stabilised using ACE inhibitors or ARB antihypertensive regimens (or both) prior to randomisation. Beta‐blockers not first‐line or monotherapy.
IMPACT‐HF 2004	Study assessed the use of carvedilol therapy initiated before discharge in people hospitalised with heart failure compared with 'usual care'. Not all participants had hypertension (64%). Baseline mean BP 124/69.5 mmHg)
MAPHY 1988	This multicentre study was a subset of the HAPPHY trial. Analysis take into consideration only 1 of the 2 beta‐blockers (metoprolol). Including this trial alongside the HAPPHY trial would count those participants twice.
Marazzi 2011	This trial compared the effects of long‐term treatment with nebivolol vs carvedilol on left ventricular ejection fraction in people with hypertensive chronic heart failure. We excluded this study because the majority of participants were already taking other antihypertensives at baseline, mainly ACE inhibitors.
MERIT‐HF 2002	Trial evaluated metoprolol compared to placebo added to standard therapy in people with heart failure. Not all participants had hypertension (44%). Mean baseline BP not given.
Nilsson 2007	This trial compared 2 first‐line antihypertensive therapies for initiating treatment in hypertension, i.e. the ACE inhibitor zofenopril and the beta‐blocker atenolol. The study has not reported effects on mortality or cardiovascular endpoints.
NORDIL 2000	The Nordic Diltiazem (NORDIL) study enrolled 10,881 people with hypertension aged 50 to 74 years at health centres in Norway and Sweden and randomly assigned them to either diltiazem, or diuretics with/without beta‐blockers. Morbidity and mortality were not reported separately for participants assigned to beta‐blocker therapy.
REASON 2009	Trial compared the effects of atenolol and perindopril/indapamide on BP and carotid‐femoral pulse wave velocity, which is a marker for aortic stiffness and arterial wall alterations. No morbidity or mortality data reported.
RESOLVD 2000	Trial compared metoprolol or placebo in people with heart failure who had received treatment with either an ACE inhibitor (enalapril) or ARB (candesartan) or both for 5 months prior to trial commencement (+ a diuretic in 84% of participants). Beta‐blocker not first‐line or monotherapy.
SENIORS 2005	Study compared the effects of nebivolol with placebo, in addition to standard therapy, in elderly people with chronic heart failure. Not all participants had hypertension (62%). Mean baseline BP 139/81 mmHg.
STOP 1991	This study compared the effects of active hypertensive treatment (1 of 3 beta‐blockers or a diuretic) and placebo in elderly people with hypertension. Morbidity and mortality were not reported separately for participants assigned to beta‐blocker therapy.
STOP‐2 1999	Conventional antihypertensive drugs (1 of 3 beta‐blockers or a diuretic) were compared with newer agents, ACE inhibitors and calcium‐channel blockers. Findings were not reported separately for participants taking beta‐blockers.
TEST 1995	The trial was conducted in 21 centres in Sweden between July 1988 and June 1992. The study evaluated the effects of a beta‐blocker (atenolol) in people aged > 40 years enrolled within 3 weeks of a stroke or transient ischaemic attack. Participants were randomised to atenolol or a matching placebo. The proportion of participants on assigned treatment at the end of the study not stated. The trial followed up 720 participants (mean age 70.4 years) for a mean duration of 2.5 years. We excluded this study because not all participants had hypertension at baseline.

ACE: angiotensin‐converting enzyme; ARB: angiotensin receptor blocker; BP: blood pressure; DBP: diastolic blood pressure; MI: myocardial infarction; RCT: randomised controlled trial; SBP: systolic blood pressure.