Effect of early treatment with anti‐hypertensive drugs on short and long‐term mortality in patients with an acute cardiovascular event

Abstract

Background

Acute cardiovascular events represent a therapeutic challenge. Blood pressure lowering drugs are commonly used and recommended in the early phase of these settings. This review analyses randomized controlled trial (RCT) evidence for this approach.

Objectives

To determine the effect of immediate and short‐term administration of anti‐hypertensive drugs on all‐cause mortality, total non‐fatal serious adverse events (SAE) and blood pressure, in patients with an acute cardiovascular event, regardless of blood pressure at the time of enrollment.

Search methods

MEDLINE, EMBASE, and Cochrane clinical trial register from Jan 1966 to February 2009 were searched. Reference lists of articles were also browsed. In case of missing information from retrieved articles, authors were contacted.

Selection criteria

Randomized controlled trials (RCTs) comparing anti‐hypertensive drug with placebo or no treatment administered to patients within 24 hours of the onset of an acute cardiovascular event.

Data collection and analysis

Two reviewers independently extracted data and assessed risk of bias. Fixed effects model with 95% confidence intervals (CI) were used. Sensitivity analyses were also conducted.

Main results

Sixty‐five RCTs (N=166,206) were included, evaluating four classes of anti‐hypertensive drugs: ACE inhibitors (12 trials), beta‐blockers (20), calcium channel blockers (18) and nitrates (18). Acute stroke was studied in 6 trials (all involving CCBs). Acute myocardial infarction was studied in 59 trials. In the latter setting immediate nitrate treatment (within 24 hours) reduced all‐cause mortality during the first 2 days (RR 0.81, 95%CI [0.74,0.89], p<0.0001). No further benefit was observed with nitrate therapy beyond this point. ACE inhibitors did not reduce mortality at 2 days (RR 0.91,95%CI [0.82, 1.00]), but did after 10 days (RR 0.93, 95%CI [0.87,0.98] p=0.01). No other blood pressure lowering drug administered as an immediate treatment or short‐term treatment produced a statistical significant mortality reduction at 2, 10 or ≥30 days. There was not enough data studying acute stroke, and there were no RCTs evaluating other acute cardiovascular events.

Authors' conclusions

Nitrates reduce mortality (4‐8 deaths prevented per 1000) at 2 days when administered within 24 hours of symptom onset of an acute myocardial infarction. No mortality benefit was seen when treatment continued beyond 48 hours. Mortality benefit of immediate treatment with ACE inhibitors post MI at 2 days did not reach statistical significance but the effect was significant at 10 days (3‐5 deaths prevented per 1000). There is good evidence for lack of a mortality benefit with immediate or short‐term treatment with beta‐blockers and calcium channel blockers for acute myocardial infarction.

Plain language summary

Effects of blood pressure lowering drugs given in the first 24 hours of heart attack or stroke

Reducing blood pressure with drugs has been a strategy used in patients suffering from an acute event in the heart or in the brain, such as heart attack or stroke. There is controversy whether these drugs should be used in the immediate period of these events, and what would be the best type of drug that renders the most benefit. This review looked at all studies where patients were randomized to one of these drugs or placebo, in this period. One class of blood pressure lowering drug, the so‐called nitrates, demonstrated reduction in mortality in patients with heart attack. For 1000 patients treated 4 to 8 deaths were prevented during the first 2 days of this acute event. The ACE‐inhibitors class also decrease mortality when continued for 10 days (3 to 5 deaths prevented per 1000). No other class of drug showed a reduction in mortality.

Summary of findings

Summary of findings for the main comparison. Immediate and short term administration of nitrates in patients with acute cardiovascular events.

Immediate and short term administration of nitrates in patients with acute cardiovascular events
Patient or population: patients with acute myocardial infarction1 Settings: hospitalized within 24 hours of the symptom onset Intervention: Nitrates Comparison: Placebo or no treatment
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo or no treatment	Nitrates
All‐cause mortality at 2 days ‐ immediate treatment Follow‐up: 0‐48 hours	Low risk population2		RR 0.81 (0.74 to 0.89)	82624 (6 studies)	⊕⊕⊕⊕ high	Highly significant benefit was achieved. 125 or 250 patients, with high or low risk, need to be treated to prevent 1 death
	20 per 1000	16 per 1000 (15 to 18)
	High risk population2
	40 per 1000	32 per 1000 (30 to 36)
All‐cause mortality at 10 days‐ short‐term treatment Follow‐up: 0‐10 days	Low risk population2		RR 0.91 (0.86 to 0.97)	78178 (6 studies)	⊕⊕⊕⊕ high	Significant benefit is demonstrated, but this is the same absolute magnitude as day 2. Thus, likely reflects the mortality benefit at day 2.
	50 per 1000	46 per 1000 (43 to 49)
	High risk population2
	80 per 1000	73 per 1000 (69 to 78)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

¹ Mortality results for this clinical condition, at the different times, and according to the different modalities of treatment (immediate and short term) displayed here, are considered the key points for this summary of findings table .The other results can be found in text. 
 ² These low and high risk values were chosen based on the second lowest, and second highest risks in the control group of the included studies

Summary of findings 2. Immediate and short term administration of ACE inhibitors in patients with acute cardiovascular events.

Immediate and short term administration of ACE inhibitors in patients with acute cardiovascular events
Patient or population: acute myocardial infarction1 Settings: hospitalized within 24 hours of the symptom onset Intervention: ACE‐inhibitors Comparison: Placebo or no treatment
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo or no treatment	ACE‐inhibitors
All‐cause mortality at 2 days ‐ immediate treatment Follow‐up: 0‐48 hours	Low risk population2		RR 0.91 (0.82 to 1)	77414 (3 studies)	⊕⊕⊕⊕ high	No statistically significant effect, but possible benefit.
	20 per 1000	18 per 1000 (16 to 20)
	High risk population2
	40 per 1000	36 per 1000 (33 to 40)
All‐cause mortality at 10 days ‐ short‐term treatment Follow‐up: 0‐10 days	Low risk population3		RR 0.93 (0.87 to 0.98)	84311 (10 studies)	⊕⊕⊕⊕ high	Significant benefit was achieved. 200 or 333 patients, with high or low risk, need to be treated for 10 days to prevent 1 death
	40 per 1000	37 per 1000 (35 to 39)
	High risk population3
	70 per 1000	65 per 1000 (61 to 69)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

¹ Mortality results for this clinical condition, at the different times, and according to the different modalities of treatment (immediate and short term) displayed here, are considered the key points for this summary of findings table . The other results can be found in text. 
 ² These low and high risk values were chosen based on the lowest and highest risks in the control group of these 3 included studies 
 ³ These low and high risk values were chosen based on the second lowest, and second highest risks in the control group of the included studies

Summary of findings 3. Immediate and short term administration of Beta‐blockers in patients with acute cardiovascular events.

Immediate and short term administration of Beta‐blockers in patients with acute cardiovascular events
Patient or population: patients with acute myocardial infarction1 Settings: hospitalized within 24 hours of symptom onset Intervention: Beta‐blockers Comparison: Placebo or no treatment
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo or no treatment	Beta‐blockers
All‐cause mortality at 2 days ‐ immediate treatment Follow‐up: 0‐48 hours	Medium risk population2		RR 0.95 (0.85 to 1.07)	68007 (6 studies)	⊕⊕⊝⊝ low3,4	No statistically significant effect.<BR/>
	15 per 1000	14 per 1000 (13 to 16)
All‐cause mortality at 10 days ‐ short‐term treatment Follow‐up: 0‐10 days	Medium risk population5		RR 0.96 (0.91 to 1.02)	71457 (14 studies)	⊕⊕⊕⊕ high	No statistically significant effect.
	42 per 1000	40 per 1000 (38 to 43)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

¹ Mortality results for this clinical condition, at the different times, and according to the different modalities of treatment (immediate and short term) displayed here, are considered the key points for this summary of findings table . The other results can be found in text 
 ² Medium risk was chosen as there was little variation in the control group risk across included trials 
 ³ Performance bias was highly suspected in a large open label‐trial 
 ⁴ There was significant variability in the effect estimate 
 ⁵ This is the medium control group risk from the included studies

Summary of findings 4. Immediate and short term administration of calcium channel blockers (CCB) in patients with acute cardiovascular events.

Immediate and short term administration of calcium channel blockers (CCB) in patients with acute cardiovascular events
Patient or population: patients with acute myocardial infarction or stroke1 Settings: hospitalized within 24 hours of the onset Intervention: CCB Comparison: Placebo or no treatment
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo or no treatment	CCB
All‐cause mortality at 2 days ‐ immediate treatment Follow‐up: 0‐48 hours	Medium risk population2		RR 2.33 (0.62 to 8.78)	242 (3 studies)	⊕⊝⊝⊝ very low3,4,5	No statistically significant effect. Not enough trials/patients or events.
	13 per 1000	30 per 1000 (8 to 114)
All‐cause mortality at 10 days ‐ short‐term treatment Follow‐up: 0‐10 days	Medium risk population6		RR 1.01 (0.73 to 1.38)	1900 (15 studies)	⊕⊝⊝⊝ very low3,4,5	No statistically significant effect. Not enough trials/patients or events.
	70 per 1000	71 per 1000 (51 to 97)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

¹ Mortality results for these clinical conditions, at the different times, and according to the different modalities of treatment (immediate and short term) displayed here, are considered the key points for this summary of findings table .The other results can be found in text 
 ² Only the medium risk was chosen as there was little variation in the control group risk across included trials 
 ³ These were very small open‐label trials 
 ⁴ The 95 % confidence intervals of the effect estimate goes in opposite direction across trials reaching the threshold for clear benefit (RR=0.75) and clear harm (RR=1.25) 
 ⁵ Probably there are many missing reports 
 ⁶ This is the median control group risk from the included studies

Background

Introduction

Acute cardiovascular events represent a significant therapeutic challenge as they are the number one cause of hospitalizations in many countries including Canada (Health Canada 2003). A wide range of pharmacological interventions administered for these conditions have been studied in great detail worldwide. However, some questions remain unanswered. For instance, it is not known whether blood pressure lowering drugs given in the early phase of these events is beneficial or harmful. The best time for initiation of treatment with these drugs is also not known, although most clinical guidelines recommend starting treatment within 24 hours of the onset. For example, the ACC/AHA, 2004 recommend to "start the administration of ACE inhibitors within the first 24 hours of an acute myocardial infarction" . In the setting of an acute cardiovascular event, in general, most of the time blood pressure is not elevated and sometimes it is low. Thus, the rationale for the early administration of blood pressure lowering drug is not completely clear. The objective of this review is to find randomized controlled trial (RCT) evidence for the use of blood pressure lowering drugs in this early phase of an acute CVE.

An acute cardiovascular event is a sudden manifestation of a cardiovascular disease (e.g. myocardial infarction or stroke). According to the World Health Organization, cardiovascular diseases are the leading cause of death in the world (WHO 2004). Acute cardiovascular events include acute myocardial infarction (AMI), unstable angina (UA), acute stroke, acute aortic dissection, and left‐ventricular failure with acute pulmonary edema. Although these could be considered different clinical entities they share one thing in common; they are life‐threatening and have a high rate of initial mortality if left untreated. For example, epidemiological studies from the 1960's showed that 75% of patients died within 24 hours of the onset of acute myocardial infarction (Fulton 1969).

The best way to understand the rationale for this review is to consider another well studied intervention, fibrinolysis which is administered in the early phase of an acute myocardial infarction or stroke. Consider the example of a large RCT (ISIS‐2 1988) where 17,187 patients within 24 hours of the onset of a suspected acute myocardial infarction were randomized to receive one‐hour streptokinase IV infusion or placebo. The patients were followed for 35 days and relative mortality between the two groups was measured over time. By inspecting the survival curves it was possible to determine that 75% of the deaths in the placebo group, occurred during the first 10 days. Streptokinase had no effect on mortality at 48 hours (RR=1.04, 95%CI[0.89,1.21]), but reduced mortality at 10 days (RR= 0.82, 95%CI[0.74,0.91]) and at 35 days (RR=0.77, 95%CI[0.70,0.84]). From this it can be deduced that this immediate and short treatment had a beneficial effect, but this was only manifest when measured at 10 days and 35 days.

This trial, thus, represents the ideal design we are looking for in this review where an early (within 24 hours) antihypertensive intervention is given for a short time (up to 48 hours) to patients with an acute cardiovascular event, and the key outcomes are reported at 48 hours, 10 days and ≥30 days. We have chosen these times as it has been emphasized that when the severity of a disease changes quickly over time, the follow‐up periods need to be subdivided into an early period, an intermediate period and a late period (Peto 1977). In addition to this ideal trial design, we have also studied and analyzed trials in which the intervention was started within 24 hours and continued for up to 10 days. We refer to the 1 to 48 hour intervention as "immediate" and the intervention started within 24 hours and continued up to 10 days as "short‐term".

For trials where the experimental treatment and control was continued more than 10 days we limited our analysis to the 2 day and 10 day mortality. Most patients are stable after 10 days and drug effects occurring after the acute period are not within the purview of this review.

Description of the intervention

Anti‐hypertensive or blood pressure lowering drugs are defined as those pharmacological agents indicated and used to treat elevated blood pressure or hypertension. According to the World Health Organization (WHO) / International Society of Hypertension (ISH) (WHO/ISH 2003) and other international guideline committees (ESH‐ESC 2007, BHS‐IV, 2004, JNC‐7, 2003) these include angiotensin converting enzyme inhibitors (ACE‐I), angiotensin II receptor blockers (ARBs), beta‐adrenergic receptor blockers (BB), calcium channel blockers (CCB), diuretics and nitrates (including nitroprusside).

It is important to emphasize that anti‐hypertensive drugs have the potential to cause pharmacological actions other than lowering BP (for example, reducing heart rate) when administered to humans, which makes them usable for indications other than treating hypertension. In contrast, there are drugs that have the potential to reduce blood pressure (for example: morphine in certain dose and setting) but are not classified as anti‐hypertensive drugs. The latter type of drugs are not considered in this review.

Anti‐hypertensive drugs are commonly used (Jackevicius 2005) and recommended in the early phase of an acute cardiovascular event. The most common anti‐hypertensive drugs that are recommended are ACEI for acute myocardial infarction (AMI) (ACC/AHA, 2004), (ESC, 2003) and for unstable angina or non‐ST elevation myocardial infarction (UA/NSTEMI) (ACC/AHA, 2007); BB for AMI (ACC/AHA, 2004, ESC, 2003), for UA/NSTEMI (ACC/AHA, 2007) and for stroke (hypertensive) (AHA/ASA, 2005), CCB for AMI (ACC/AHA, 2004) and for Stroke (hypertensive) (AHA/ASA, 2005), diuretics for AMI/APE (ACC/AHA, 2004), and nitrates for AMI (ACC/AHA, 2004, ESC, 2003), for UA/NSTEMI (ACC/AHA, 2007) and for Stroke (hypertensive) (AHA/ASA, 2005).

How the intervention might work

The exact mechanism of action of blood pressure lowering drugs, antihypertensives, is often not known with certainty, but each of the different classes of drugs act at different sites and by different mechanisms.

Angiotensin converting enzyme inhibitors (ACE‐I)

ACE‐I inhibit the conversion of angiotensin I, a relatively inactive peptide, to angiotensin II, which is a potent vasoconstrictor and causes release of aldosterone from the adrenal cortex. The ACE‐I probably lower blood pressure by reducing the blood pressure increasing effects of angiotensin II.

Angiotensin II receptor blockers (ARBs)

ARBs inhibit the binding and action of angiotensin II at the angiotensin II type 1 (AT1) receptor, which is the site of action to cause vasoconstriction and release of aldosterone. They, thus, also probably reduce BP by blocking the BP increasing effects of angiotensin II.

Beta‐adrenergic receptor blockers (BB)

In general beta‐adrenergic receptor blocking drugs or beta‐blockers (BB) are drugs that competitively inhibit the effect of the catecholamines, noradrenaline and adrenaline, on beta‐adrenergic receptors. There are subtypes of beta‐blockers depending on their relative affinity for B1 and B2 receptors, partial agonist activity, and ability to also block alpha‐adrenergic receptors. Catecholamines have a positive chronotropic and inotropic actions on the heart, can cause vasoconstriction or vasodilatation of blood vessels and stimulate renin release from the kidney. Thus, BB drugs could lower BP by a number of these differing effects. The exact mechanism by which beta‐blockers lower BP in humans is not known.

Calcium channel blockers (CCB)

Calcium channel blockers reduce the cytosolic free‐calcium concentrations by blocking transmembrane calcium influx through L‐type calcium channels. Dihydropyridines (such as nifedipine), benzothiazepines (diltiazem) and phenylalkylamines (verapamil) bind to the pore‐containing the α1 subunit of the L‐type calcium channel. In general, calcium channel blockers relax arteriolar smooth muscle, resulting in vasodilatation and decreased peripheral resistance.  The decreased systemic resistance is thought to cause the blood pressure reduction. Agents that slow the rate of recovery of L‐type calcium channels (verapamil, diltiazem) have negative chronotropic and dromotropic effects on the heart's conducting system. The CCBs also have a natriuretic effect on the kidney that may contribute to their ability to lower blood pressure.

Diuretics

Thiazide diuretics inhibit the Na⁺Cl^‐ co‐transporter in the proximal part of the distal convoluted tubule of the kidney. This decreases tubular reabsorption of sodium and chloride, and increases urinary excretion of sodium and water .

Furosemide and other loop diuretics inhibit sodium reabsorption in the ascending limb of loop of Henle as well as in both proximal and distal tubules.

The mechanism by which diuretics lower blood pressure has not been established.

Nitrates (including nitroprusside)

Organic nitrates and sodium nitroprusside (SNP) are nitric oxide (NO) donors. Organic nitrates, such as glyceryl trinitrate (nitroglycerine) require enzymatic metabolism to generate NO. In contrast SNP spontaneously generate this NO. There are a number of theories about how nitrates cause vasodilation: 1) acting on specific nitrate receptor (containing a sulfhydryl group‐SH), 2) NO exerts a potassium channel activation (hyperpolarizing the cell membrane) and 3) NO activates the enzyme guanylate cyclase increasing cGMP levels, which in turn inhibits Ca⁺ entry into smooth muscle cells and increases Ca⁺ uptake by the smooth endoplasmic reticulum resulting in vascular smooth muscle relaxation.

Why it is important to do this review

First, to determine whether blood pressure lowering by all classes of blood pressure lowering drugs during the early phase of an acute cardiovascular event affects mortality and morbidity.

It is important to appreciate that during the first hours after an acute cardiovascular event, complex hemodynamic changes are occurring, making the organs involved especially vulnerable to local and systemic changes. Blood pressure reduction during this time could be beneficial, but could equally likely be detrimental.

Recently, thanks to the advances in the understanding of the underlying pathophysiological mechanisms involved in some acute cardiovascular events (such as the thrombotic occlusion at the site of atherosclerotic plaques in AMI) different pharmacological interventions have been implemented and proven beneficial. For example, randomized controlled trial (RCT) evidence has demonstrated a reduction in mortality with the use of thrombolytic drugs (FTT 1994, Boersma 1996) and anti‐platelet drugs (ATC 2002) in acute AMI. There is also RCT evidence for benefit with the use of antiplatelet drugs in the acute treatment of stroke (ATC 2002). The benefits of these pharmacological therapies have been claimed to be greater when they are administrated in the early phase (for example: thrombolysis within 6 hours) of the onset of the acute cardiovascular event. Thus, the time of the administration can be an important determinant as to whether an intervention is beneficial or harmful in patients with an acute cardiovascular event.

The rationale for administering blood pressure lowering drugs in the early phase of an acute cardiovascular event is less well understood. In acute cardiovascular events associated with marked elevation of blood pressure (defined as hypertensive emergencies by JNC‐7, 2003) blood pressure lowering makes some pathophysiological sense as if the elevated blood pressure is causing organ damage, reducing blood pressure is likely to be beneficial. However, there is no RCT evidence to demonstrate mortality and morbidity benefits of the use of blood pressure lowering drugs in hypertensive emergencies (Perez 2008).

This lack of evidence for situations where blood pressure is elevated emphasizes the importance of attempting to assess the benefits and harms in acute cardiovascular events in general. In the setting of acute cardiovascular event in general most of the time blood pressure is not elevated and sometimes it is low. 

We have deliberately chosen outcomes in this review that are a measure of both benefit and harm in order to avoid selective reporting bias. Due to the standardized definition of serious adverse events (SAE), i.e. any untoward medical occurrence that results in death, is life‐threatening, requires hospitalization or prolongation of hospitalization, or results in persistent or significant disability (ICH/FDA 1995), measuring total people with at least one serious adverse event would provide a very close estimate of the net health effect (benefit minus harm) of an intervention (Wright 2002). However, in acute, critically‐ill hospitalized patient settings measuring SAE could be problematic and/or impractical. The main reason for this is because the physician must make a judgment as to whether an event led to prolongation of hospitalization. Therefore, in these acute clinical settings total all‐cause mortality, a subgroup of total SAEs, is the best measure of net health effect (Cook 2008). All‐cause mortality is an outcome measure that is not subject to physician judgment and which is usually reported in trials. It is easy to appreciate that an intervention that decreases mortality as compared to placebo is beneficial, while an intervention that increases mortality compared to placebo is harmful. Thus, we have chosen all‐cause mortality as our primary outcome and total non‐fatal serious adverse event as one of our secondary outcomes anticipating that trials might not consistently report this outcome. 

Secondly to determine whether blood pressure lowering by the different subclasses of blood pressure lowering drugs during the early phase of an acute cardiovascular event affects mortality and morbidity. This question can be answered by performing a subgroup analysis where the effect of the drugs in the different classes of blood pressure lowering drugs are analyzed separately and compared with each other.

There are other published systematic reviews of randomized controlled trials dealing with antihypertensive drugs for acute cardiovascular events (specifically for acute myocardial infarction or stroke). For acute myocardial infarction reviews have been published for each different class of anti‐hypertensive drug: CCB (Held 1989); BB (Yusuf 1985, Freemantle 1999); nitrates (Yusuf 1988), and ACE inhibitors (AMICG 1998, Rodrigues 2003). In addition there are three published systematic reviews dealing with acute stroke: CCB (Horn 2000), nitrates (Bath 2002), and diverse interventions (Geeganage 2008).

These systematic reviews all have significant drawbacks and none were designed to assess the specific objective of this review. Some were not limited to immediate treatment (Rodrigues 2003, Al‐Reesi 2008, AMICG 1998, Held 1989, Yusuf 1985, Bath 2002, Freemantle 1999) and included trials where the treatment started days after the onset of the acute cardiovascular event. One review mixed trials with control group (no treatment or placebo) with trials that have an active comparator as control (Yusuf 1988). Except for Geeganage 2008, all these reviews are also limited by not being up‐to‐date.

This is the first systematic review with the objective to assess the effects of all blood pressure lowering drugs administered as immediate treatment (starting within 24 hours) in patients with an acute cardiovascular event, on mortality, morbidity and blood pressure outcomes.

Objectives

Primary

To determine the effect of immediate^ and short‐term^^ treatment with antihypertensive drugs on mortality at 2 days, 10 days and ≥ 30 days in patients with an acute cardiovascular event.

^ Immediate treatment is defined as treatment started within 24 hours of the onset of an acute cardiovascular event and lasting for a maximum of 2 days.

^^Short‐term treatment is defined as treatment started within 24 hours of the onset of an acute cardiovascular event and lasting for a maximum of 10 days.

Secondary

To determine the effect of anti‐hypertensive drugs on blood pressure and heart rate during the first 24 hours of treatment in patients with an acute cardiovascular event.

Methods

Criteria for considering studies for this review

Types of studies

• Randomized controlled trials (RCTs) with parallel design comparing an anti‐hypertensive drug with placebo or no treatment in patients with an acute cardiovascular event.

• The intervention (anti‐hypertensive drug) must be started within 24 hours of the onset of the acute cardiovascular event.

• Patient must be followed for at least 24 hours and mortality or SAE data must be provided at least one of the specified time periods, 2 days, 10 days or ≥ 30 days.

Types of participants

Participants with any of the following acute cardiovascular events: myocardial infarction, unstable angina, acute left‐ventricular failure with pulmonary oedema, acute aortic dissection, stroke, intracranial haemorrhage, sub‐arachnoid haemorrhage.

Types of interventions

• Intervention: Any anti‐hypertensive drug†

• Control: Placebo or no early antihypertensive treatment *

†Anti‐hypertensive drug belonging to any of the following classes of drugs: nitrates (including nitroprusside), beta adrenergic antagonists (BB), angiotensin converting enzyme inhibitors (ACE‐I), calcium channel blockers (CCB), dopamine agonists, alpha‐adrenergic antagonists, diuretics (furosemide and thiazides), direct vasodilators (diazoxide, hydralazine) and others (reserpine, clonidine, alfa‐methyldopa, trimethaphan).

* Placebo is defined as inert substance designed to resemble the drug being tested but which has no active ingredient and has no treatment effect. In trials, where a placebo is used as a comparator, all patients in the placebo group usually receive the same medical treatment, except for the drug being tested, as the experimental group. This is achieved by the utilization of a double‐blind study design in these trials. In trials where no treatment is used as a comparator it is assumed that all medical treatments other than antihypertensive treatment intervention being studied are the same in both groups. In these trials the design is open‐label and not blinded.

Types of outcome measures

Primary outcomes

The primary outcome is all‐cause mortality at 2 days, 10 days and ≥ 30 days.

Operational definitions:

• At 2 days also accepts cumulative mortality at times less than 48 hours if that is the only data available.

• At 10 days also accepts cumulative mortality reported after 2 days and at times less than 10 days if that is the only data available.

• At ≥ 30 days also accepts cumulative mortality reported as 1‐month or 4‐week mortality and for any duration of follow up longer than 30 days.

Secondary outcomes

• Total number of patients with at least one non‐fatal Serious Adverse Event (SAE) at the same time periods for all cause mortality.

• Weighted mean change in systolic blood pressure (SBP), diastolic blood pressure (DBP), and heart rate (HR), during the first 24 hours of treatment.

Search methods for identification of studies

Electronic searches

The Database of Abstracts of Reviews of Effectiveness (DARE) and the Cochrane Database of Systematic Reviews were searched for related reviews.

The following electronic databases were searched for primary studies:

1. The Cochrane Central Register of Controlled Trials (CENTRAL)

2. English language databases, including MEDLINE (1966‐February 2009) and EMBASE (1982‐February 2009)

Electronic databases were searched using a strategy combining a variation of the Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE: sensitivity‐maximizing version (2008 revision) with selected MeSH terms and free text terms relating to antihypertensives and acute cardiovascular events.  No language restrictions were used.  The MEDLINE search strategy was translated into the other databases using the appropriate controlled vocabulary as applicable. 

Full electronic database search strategies are included in Appendix 1.

Searching other resources

Reference lists of all papers and relevant reviews were identified, as well as reference lists of review articles.

Authors of trials reporting incomplete information were contacted to provide the missing information.

Data collection and analysis

Data extraction: 
 Two reviewers (MIP & VM) independently decided whether a trial was included. They also extracted and verified data entry from included studies. Discrepancies were resolved by discussion. Absence of consensus was resolved by a third reviewer (JMW).

Analyses: 
 For the analysis of the data, Cochrane Review Manager software, RevMan 5, was used. 
 Quantitative analyses of outcomes were based on intention‐to‐treat principles as much as possible. We used weighted mean difference to combine continuous variables and and expressed relative and absolute risk difference (with 95% confidence interval) for dichotomous outcomes to accept significant differences. Pooled risk differences obtained from a DerSimonian‐Laird fixed effect model were converted to numbers needed to treat (NNTs) where appropriate. 
 Heterogeneity between trial results was tested using the I² statistic where percentages greater than 50% were taken to indicate significant heterogeneity.

The effects of immediate treatment with antihypertensive drugs on 2 day, 10 day and ≥30 day mortality in patients with an acute cardiovascular event were explored. The effects of short‐term treatment with antihypertensive treatment was explored at 10 days and ≥30 days. In trials where the study treatment was continued longer than day 10, data was not included for the ≥30 day time period. To quantify the occurrences of deaths the cumulative incidence reported in each trial (based on full Intention to treat principles) was used. For example, for our day 10 mortality measure we included the cumulative incidence from time of randomization to day 10 inclusive (or at discharge from hospital when trial did not report the time ‐ the assumption here is that the average hospitalization is about 7‐10 days).

A sensitivity analysis was performed to test the impact of the design of the trial, double‐blind vs. open label trials. Sensitivity analysis were also performed according to the class of drug, clinical condition, dose regimen, duration of treatment and concomitant pharmacological interventions.

Non‐fatal serious adverse events (NF‐SAEs) were included only if they were reported as total, not as individual non‐fatal SAE. The reason for doing this was to avoid double counting individual NF‐SAE (as one patient may suffer from one or more NF‐SAE and could be included more than once in the original report) and to prevent selection reporting bias (as many trials could omit certain NF‐SAEs).

Data for blood pressure and heart rate were combined using a weighted mean difference method, whereby the trials are weighted according to the number of subjects in the trial and the within‐study variance. Some of the trials did not report a within‐study variance for blood pressure change; in these studies standard deviation (SD) of change was imputed using the following hierarchy :

1. Pooled standard deviation calculated either from the t‐statistic corresponding to an exact p‐value reported or from the 95% confidence interval of the mean difference between treatment group and comparative group.

2. Standard deviation of blood pressure/heart rate at the end of treatment.

3. Standard deviation of blood pressure/heart rate at baseline (except if this measure is used for entry criteria).

4. Weighted mean standard deviation of change in blood pressure/heart rate calculated from at least 3 other trials using the same drug and dose regimen.

5. Weighted mean standard deviation of change in blood pressure/heart rate calculated from other trials using the same drug.

6. Weighted mean standard deviation of change in blood pressure/heart rate calculated from all other trials (any drug and dose).

7. Weighted mean standard deviation in blood pressure/heart rate at end of treatment calculated from at least 3 other trials using the same drug and dose regimen.

8. Weighted mean standard deviation in blood pressure/heart rate at end of treatment calculated from all other trials (any drug and dose).

Results

Description of studies

Results of the search

The search strategy yielded 3412 citations and 82% of these were excluded after reading titles and abstracts (Figure 1). The remaining 623 citations were retrieved for detailed evaluation and 410 were excluded for the reasons shown in Figure 1. Of the 213 citations of randomized controlled trials we identified, 102 (65 trials) were included in this review. The specific reasons for exclusion of the other 111 (70 trials) are explained in the Characteristics of excluded studies table. Most of these trials had to be excluded because they did not report mortality data.

1.

Included studies

Sixty‐five randomized controlled trials (N=166,206) were found that satisfied the inclusion criteria. Of those, 40 (60%) were double‐blind placebo‐controlled trials, involving 75% (N=125,487) of the entire studied population (See Table 5). All included trials reported mortality for at least one of the time points of interest and none of the trials reported the total number of non‐fatal serious adverse events. The largest trial included 58,050 patients (ISIS‐4 1995). Twenty one trials studied immediate treatment and 44 trials studied short‐term treatment. The duration of treatment was for more than 10 days in 27/65 trials. We were able to extract short‐term mortality (at day 10) in 62 trials. Four classes of blood pressure lowering drugs were evaluated: ACE inhibitors (12 trials), B‐blockers (20 trials), Calcium channel blockers (18 trials) and nitrates, including nitroprusside (18 trials). The included trials studied patients with only 2 types of cardiovascular events: acute myocardial infarction (59 trials) and stroke (6 trials).

1. Overview of the included trials according to treatment group.

	Active drug	Placebo or No treatment
Participants in trials* involving (No.of trials):
ACE inhibitors(12)	42,260	42,196
B‐blockers(20)	36,338	36,262
CCB(18)	1,111	1,030
Nitrates(18)	42,206	42,207
All trials (65)^	83,234^	82,972^
Participants in trial design (No. of trials):
Double‐blind (40)	62,849(76%)	62,638(75%)
Not double‐blind (25)	20,385 (24%)	20,334 (25 %)
Weighted mean age (years) [trials / participants with data]	60.91 [49 / 43,243]	60.85 [49/ 43,000]
Weighted mean blood pressure (mm Hg) @ baseline [trials / participants with data]
Systolic	133.7 [34 / 41,862]	133.7 [34 / 41,669]
Diastolic	82.2 [25 / 7,125]	82.7 [25 / 7,043]
Female patients (%) [Based on trials reporting gender]	26% [47 / 82,342]	25% [47 / 82,036]
//////////////////////////////////////////////////////////////////////////////	//////////////////////////////////	////////////////////////////////////

* See below references of the randomized controlled trials for each row.

^ This total is less than the sum of all four categories as 2 trials (GSSI‐3, ISIS‐4) used 2x2 factorial design and one trial (Hargreaves 1992) compare drugs from two different categories of anti‐hypertensive with the same placebo or control group.

The link to the references of the RCTs according to each row is as follow:

ACEi (12 trials):Bussmann 1992, CONSENSUS‐II 1992, Di Pasquale 1994, Di Pasquale 1997, PRACTICAL 1994, Galcera 1993, GISSI‐3 1994, Hargreaves 1992, ISIS‐4 1995, Nabel 1991, Schulman 1995, Wagner 2002.

B‐Blockers (20 trials):

Clausen 1966, COMMIT 2005, Heber 1987, ICSG 1984, ISIS‐1 1986,Johannessen 1987; MIAMI 1985; MILIS 1984; Mitchell 2002; Norris 1978; Norris 1980; Norris 1984; Owensby 1985; Peter 1978; Salathia 1985; TIMI‐IIB 1991; Tonkin 1981; Van‐de 1993; von Essen 1982; Yusuf 1983.

Calcium Channel Blockers (18 trials):Branagan 1986; Crea 1985; Eichler 1985; Erbel 1988; Marangelli 2000; Muller 1984; Natale 1999; Pimenta 1985; Pizzetti 2001; Sirnes 1984; Theroux 1998; Zannad 1988; Gelmers 1988; VENUS 2001; Infeld 1999; Limburg 1990; Paci 1989; INWEST 1994.

Nitrates(18 trials):Beaufils 1988; Bussmann 1981; Charvat 1990; Chiche 1979; Cohn 1982; Durrer 1982; ESPRIM 1994; Fitzgerald 1990; Flaherty 1983; GISSI‐3 1994; Hargreaves 1992; Hildebrandt 1992; ISIS‐4 1995; Jaffe 1983; Jugdutt 1988; Jugdutt 1983; Lis 1984; Zharov 1991

All included trials enrolled their patients within 24 hours of the onset of an acute cardiovascular event (CVE). In this review , we have assumed that enrollment, randomization and administration of the study drug occurred simultaneously. The details of the process involved was not described in any of the trial reports. The objective for accepting these types of trials was to evaluate the effects of early anti‐hypertensive treatment in an acute setting. In the acute myocardial infarction (AMI) trials the average time of initiation of the study drug was 6.35 ± 4.74 hours after the onset. No stroke trial reported time of starting treatment. However, the included AMI studies scheduled their randomization 5 hours earlier than the stroke studies (12.6 ± 7.8 and 17 ± 8, respectively). We did find trials dealing with acute pulmonary edema and acute aortic dissection where patients were randomized within 24 hours. However, these trials had to be excluded for different reasons (see below excluded studies).

High blood pressure at baseline was not part of the inclusion criteria in any of 65 included trials. However, most of the trials explicitly stated that patients had to be free from shock or had a minimum BP (in general SBP of 90‐110 mmHg and/or DBP of 50‐60 mmHg) to enter the trial. For the included trials the overall weighted mean blood pressure at baseline was 133/82.2 mmHg for those randomized to an active drug and 133/ 82.7 mmHg for those randomized to the placebo or no treatment group. One quarter of patients were female and the overall weighted mean age was 60.85 years (Table 5).

Of the 22 trials in which we requested additional information, only 4 trialists responded to our request and provided additional information.

Blood pressure changes during the first 24 hours of treatment was reported in 19 trials (N=3,261). However, none of the trials reported on standard deviation of the change. Thus, this measure of variability was imputed from: the end point standard deviation (second‐choice in our pre‐specified hierarchy) in 14 trials; from the baseline value (third‐choice) in 1 trial; and from the weighted mean standard deviation at end point (last‐choice) in 4 trials.

Excluded studies

We excluded 70 trials. Although these trials could have been excluded for one or more reasons, the main one is listed below:

1. Mortality not reported or not reported in a form or time period that we could use: 47 trials

2. Uncertain time of patient inclusion following the acute event: 18 trials

3. Follow‐up of less than 24 hours: 5 trials

Forty‐seven trials were excluded due to lack of mortality data or mortality data that were not reported in a form or time period that we could use. There were three sub‐types: A) Not reporting mortality at all, in either original, additional publications or after being requested to provide such data (18 trials). B) Reporting mortality but not according to our pre‐specified time‐frame (22 trials); or reporting it in an abstract but not confirmed by additional publications or after being requested it (2 trials) ; C) Reporting mortality but not based on intention to treat (ITT) principles (7 trials). An example of the latter sub‐category is the multicenter double‐blind trial Walker 1988. Mortality rate was reported as 7/106 in the nifedipine group and 7/120 in the placebo group. After detailed reading of Walker's original and additional publications, it was found that the actual number of randomized patients was 217 in each group. Thus, the investigators failed to report mortality in approximately 50% of the randomized patients. If the information of this trial was included in our review, it would be misleading as this population represents a selected sub‐group of patients. Other trials that fell into this category were Eveson 2007, Szczechowski 1994, DAVIT I 1984, HINT 1986 , Barber 1976 and Balcon 1966.

Our justification for excluding those trials which had an uncertain time of patient inclusion following the acute event was to avoid having a heterogenous population that could lead to erroneous conclusions. The uncertain time inclusion applies to those trials that failed to explicitly state a specific time for inclusion; or trials that included their patients "on admission" or "immediately after admission". We believe this was too imprecise for determining the exact time that had elapsed from symptom onset to the inclusion of the patient and therefore to the initiation of therapy. It is possible that some patients could have had a delayed admission (symptoms began >24 hours). This also would likely bias for survivors. In addition, having a delayed admission was more common in trials conducted before 1980‐90's when infrastructure for emergency response to assist patients with acute cardiovascular event was not as uniformly developed as it is today. An illustrative example for this category of excluded trials is that of a double‐blind trial where patients were to be included "immediately after admission" in 16 Danish centres (DAVIT I 1984). It was determined that the treatment was started after 24 hours in >15% of patients with AMI; and mortality data was not reported separately for these patients.

Therefore, data from 65 trials (those excluded due to uncertain timing and those due to improper or non‐extractable mortality data) out of 70 excluded trials could potentially be added to our review. Of these 65 excluded trials, 14 involved an ACE inhibitor (N=4,407); 21 trials involved a Beta‐adrenergic Blocker (N=2,524); 20 trials involved CCBs (N=10,958);and 8 trials involved a nitrate (N=1,475). If these excluded patients were added to the included patients it would have only a small impact on the ACE inhibitor data (increase by 5.2%), BB data (increase by 3.5%), and nitrate data (increase by 1.7%). However, it could have a large impact on the calcium channel blocker data (increase ˜600%, from 2,141 to 13,099). There were 3/65 studies (N=126) that involved other classes of blood pressure lowering drugs: 2 trials for the alpha‐1 adrenergic antagonist prazosin (Kolettis 1983, Murdock 1990) and one trial for the alpha‐2 centrally acting agonist, clonidine (Manolis 1999).

Of 70 excluded studies, 52 trials involved patients with AMI, 9 trials included patients with acute stroke, 6 trials unstable angina, 3 trials acute pulmonary edema, and 1 trial sub‐arachnoid hemorrhage patients.

There was no trial identified that included patients with acute aortic dissection and compared an active drug vs. placebo or no treatment within 24 hours of the onset. To the best of our knowledge there is only one randomized controlled trial that has included exclusively patients with acute aortic dissection and randomized their patients within 24 hour of the onset (Yoshida 1998). However, this trial compared two active treatments, with no placebo arm.

Risk of bias in included studies

Forty trials (60%) of the 65 included studies were double‐blind, involving 75% (N=125,487) of the entire studied population. Less than 50% of the included trials reported adequate sequence generation, and only 23 trials (35%) reported concealment of allocation (Figure 2 and Figure 3). Although, the number of trials reporting adequate concealment of allocation was relatively small [13/40(32%) for double‐blind and 10/25(40%) for open‐label design trials] the number of patients randomized with adequate concealment of allocation was very high (121,618 patients out of the total 125,487 in double‐blind trials and 37,055 out of the total 40,719 in open‐label trials); in other words the trials that did not report concealment of allocation were small in sample size. Thus, this type of risk of bias is considered to be low in this review. Similarly, although more than 75% of trials had an incomplete reporting or did not report on immediate all‐cause mortality (Figure 2 and Figure 3), the studies that reported it were large enough (N=131,603) to keep this risk of incomplete outcome bias to a minimum. That was not the case for the ≥ 30 day mortality in which less than 25 % reported mortality at times equal or greater than day 30 and the trials that reported were small, therefore there is a significant risk of bias for mortality data at this time. Selective outcome reporting bias was not an issue in this review by limiting the inclusion of trials where our primary outcome mortality is reported.

2.

Methodological quality graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.

3.

Methodological quality summary: review authors' judgements about each methodological quality item for each included study.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4

COMPARISONS ACCORDING TO OUTCOMES

Primary Outcome: All‐cause mortality

Foundation: Twelve trials (N=152,029) were designed for or had the power to detect differences in mortality: CONSENSUS‐II 1992; GISSI‐3 1994; ISIS‐4 1995, COMMIT 2005; ISIS‐1 1986; MIAMI 1985; Norris 1984; Salathia 1985; VENUS 2001; Cohn 1982; Durrer 1982; ESPRIM 1994. Any trial reporting mortality, regardless of its design and size, was included in our meta‐analysis. We were able to assess the effects of immediate or short‐term treatment with 4 antihypertensive classes of drugs: ACE inhibitors, B‐blockers, calcium channel blockers and nitrates, on all‐cause mortality data at 2, 10 or ≥ 30 days; in patients presenting with acute myocardial infarction or stroke.

NITRATES

Immediate treatment ( started within 24 hours of the onset and lasting for maximum 2 days)

All‐cause mortality at 2 days (18 eligible trials, N=84,413):

Six trials (N=82,624) reported mortality at day 2. Nitrates were associated with a statistically significant reduction in all‐cause mortality (RR 0.81, 95%CI [0.74,0.89], p<0.0001;I²=0%). Figure 4

4.

Forest plot of comparison: 4 Nitrates, outcome: 4.1 All‐cause mortality at 2 days.

All‐cause mortality at 10 days (10 eligible trials, N=6,007):

All 10 trials (N=6,007) reported mortality at day 10. Nitrates were not associated with a statistically significant delayed reduction in all‐cause mortality (RR 0.84, 95%CI[0.69,1.01],p=0.07, I²=63%). Figure 5

5.

Forest plot of comparison: 4 Nitrates, outcome: 4.2 All‐cause mortality at 10 days.

All‐cause mortality at ≥ 30 days (10 eligible trials, N=6,007):

Seven trials (N=5,771) reported mortality at ≥30 days. Nitrates were not associated with a statistically significant delayed reduction in all‐cause mortality (RR 0.92, 95%CI[0.82,1.04],p=0.20, I²=42%) during a weighted average of 12 months of follow‐up. Figure 6

6.

Forest plot of comparison: 4 Nitrates, outcome: 4.3 All‐cause mortality at ≥30 days.

Short‐term treatment ( started within 24 hours of the onset and lasting for a maximum of 10 days)

All‐cause mortality at 10 days: (8 eligible trials, N=78,406):

Six trials (N=78,178) reported mortality at 10 days. Nitrates were associated with a statistically significant reduction in all‐cause mortality (RR 0.91,95%CI [0.86,0.97], p=0.003; I²=0%) Figure 5.

All‐cause mortality at ≥ 30 days (5 eligible trials, N=969):

Three trials (N=570) reported mortality at ≥30 days. Nitrates were not associated with a statistically significant delayed reduction in all‐cause mortality (RR 0.72, 95%CI [0.48,1.10],p=0.13, I²=81%) during a weighted average of 4 months of follow‐up Figure 6.

ANGIOTENSIN CONVERTING ENZYME (ACE) INHIBITORS

Immediate treatment ( started within 24 hours of the onset and lasting for maximum 2 days)

All‐cause mortality at 2 days: (12 eligible trials, N=84,456):

Three trials (N=77,414) reported mortality at 2 days. ACE inhibitors were not associated with statistically significant reduction in all‐cause mortality (RR 0.91, 95%CI [0.82,1.00]), p=0.05, I²=0%] Figure 7.

7.

Forest plot of comparison: 1 ACEi, outcome: 1.1 All‐cause mortality at 2 days.

All‐cause mortality at 10 days (2 eligible trials, N=145,Bussmann 1992; Wagner 2002):

The 2 trials (N=145) reported mortality at 10 days. ACE inhibitors were not associated with a statistically significant delayed reduction in mortality at 10 days (RR 0.68, 95%CI[0.12,3.98],p=0.67).

All‐cause mortality at ≥ 30 days (2 eligible trials, N=145):

No trial reported mortality at this time.

Short‐term treatment ( started within 24 hours of the onset and lasting for a maximum of 10 days)

All‐cause mortality at 10 days (10 eligible trials, N=84,311):

All 10 trials (N=84,311) reported mortality at 10 days. ACE inhibitors were associated with a statistically significant reduction in all cause mortality as compared to placebo (RR 0.93, 95%CI [0.87,0.98] p=0.01). Figure 8

8.

Forest plot of comparison: 1 ACEi, outcome: 1.2 All‐cause mortality at 10 days.

All‐cause mortality at ≥ 30 days (2 eligible trials, N=432,Di Pasquale 1994,Di Pasquale 1997):

No trial reported mortality at ≥ 30 days .

BETA‐ADRENERGIC ANTAGONIST, BETA‐BLOCKERS (BB)

Immediate treatment ( started within 24 hours of the onset and lasting for maximum 2 days)

All‐cause mortality at 2 days(20 eligible trials, N=72,600):

Six trials (N=68,007) reported mortality at 2 days. BB drugs were not associated with a statistically significant reduction in all‐cause mortality (RR 0.95, 95%CI [0.85,1.07],p=0.39, I²=67% ) Figure 9.

9.

Forest plot of comparison: 2 BB, outcome: 2.1 All‐cause mortality at 2 days.

All‐cause mortality at 10 days (6 eligible trials, N=1,143):

All 6 trials (N=1,143) reported mortality at 10 days. BB drugs were not associated with a statistically significant delayed reduction in all‐cause mortality (RR 1.12, 95%CI [0.60,2.07],p=0.73, I²=0%) Figure 10

10.

Forest plot of comparison: 2 BB, outcome: 2.3 All‐cause mortality at 10 days.

All‐cause mortality at ≥ 30 days (6 eligible trials, N=1,143):

Only one trial (N=108) reported mortality at ≥30 days. There were no deaths in either group.

Short‐term treatment ( started within 24 hours of the onset and lasting for a maximum of 10 days)

All‐cause mortality at 10 days (14 eligible trials, N=71,457):

All 14 trials (N=71,457) reported mortality at 10 days. BB drugs were not associated with a statistically significant reduction in all‐cause mortality (RR 0.96, 95%CI[0.91,1.02],p=0.21, I²=0%) Figure 10.

All‐cause mortality at ≥ 30 days (8 eligible trials, N=18,645):

Five trials (N=18,373) reported mortality at ≥ 30 days. BB drugs was not associated with a statistically significant delayed reduction in all‐cause mortality (RR 0.91,95%CI [0.84,0.99], p=0.03, I²=0%) during a weighted average of 12 months of follow‐up Figure 11.

11.

Forest plot of comparison: 2 BB, outcome: 2.4 All‐cause mortality at ≥30 day.

CALCIUM CHANNEL BLOCKERS (CCB)

Immediate treatment ( started within 24 hours of the onset and lasting for maximum 2 days)

All‐cause mortality at 2 days (18 eligible trials, N=2,141):

Only 3 trials (N=242, Erbel 1988;Theroux 1998;Zannad 1988) reported mortality at 2 days. CCB were not associated with a statistically significant reduction in all‐cause mortality (RR 2.33,95%CI [0.62,8.78],p=0.21, I²=2%) Figure 12.

12.

Forest plot of comparison: 3 CCB, outcome: 3.1 All‐cause mortality at 2 days.

All‐cause mortality at 10 days (3 eligible trials, N=241,Branagan 1986, Eichler 1985, Marangelli 2000):

Only one trial Marangelli 2000 (N=88) reported mortality at 10 days. There was one death in the CCB group and none in the control group.

All‐cause mortality at ≥ 30 days (3 eligible trials, N=241):

Only one trial Branagan 1986 (N=108) reported mortality at ≥ 30 days. There were 7/54 deaths in the CCB group and 5/54 in the control group. The relative risk for this trial was 1.40, 95%CI [0.47,4.14].

Short‐term treatment ( started within 24 hours of the onset and lasting for a maximum of 10 days)

All‐cause mortality at 10 days (15 eligible trials, N=1,900):

All 15 trials (N=1,900) reported mortality at 10 days. CCB were not associated with a statistically significant effect on all‐cause mortality (RR 1.01, 95%CI[0.73,1.38], I²=0%) Figure 13.

13.

Forest plot of comparison: 3 CCB, outcome: 3.2 All‐cause mortality at 10 days.

All‐cause mortality at ≥ 30 days (5 eligible trials, N=730):

One trial (N=90), Pizzetti 2001 reported mortality at ≥ 30 days. There were three deaths in the CCB group and two in the placebo group.

SECONDARY OUTCOME: Total non‐fatal serious adverse events:

No trial reported total non‐fatal serious adverse events (SAE). it was not possible to extract individual non‐fatal SAE from the original trials and analyze them as a composite due to a risk of double‐counting the events and due to the risk of missing particular non‐fatal SAEs that were not reported.

SECONDARY OUTCOME: Weighted mean change in Blood pressure and Heart rate during the first 24 hours of treatment:

Many trials did not report these secondary outcomes. In general, trials that reported blood pressure data and also mortality data were small trials (average 170 patients). We carefully looked at all trials to see if they have blood pressure data for the entire population or for a subset of the entire study. Cohn 1982 was the only trial designed to detect mortality differences and that reported blood pressure data. The MIAMI 1985 trial had a subset study with blood pressure data and was included in our meta‐analysis. It was not possible to do a secondary analysis (meta‐regression or correlation ) comparing the effects of blood pressure lowering during the first 24 hours and health outcomes. No trial reported the standard deviation of the change for blood pressure or heart rate. Thus, this measure of variability was imputed from: the end point (second‐top choice in our pre‐specified hierarchy) in 14 trials; from the baseline value (third choice) in 1 trial; and from the weighted mean standard deviation at end point (last choice) in 4 trials.

Angiotensin converting enzyme inhibitors (ACEi)

Systolic blood pressure change: There were no included trials providing SBP data for ACEi.

Diastolic blood pressure change: There were no included trials providing DBP data for ACEi.

Heart rate change: There were no included trials providing HR data for ACEi.

Beta‐adrenergic antagonist or Beta‐Blockers (BB)

Systolic blood pressure change: There were 6 trials (N=738) included in the systolic meta‐analysis. The pooled effect showed a statistically significant greater reduction in SBP with beta‐blockers drugs compared to placebo or no treatment during the first 24 hours of a cardiovascular event (WMD ‐12.54, 95%CI [‐15.63,‐9.45], p<0.00001;I²=0%)Figure 14.

14.

Forest plot of comparison: 2 BB, outcome: 2.4 Weighted mean change in Systolic Blood pressure during first 24 hours.

Diastolic blood pressure change: There were 6 trials (N=738) included in the diastolic meta‐analysis. The pooled effect showed a statistically significant greater reduction in DBP with beta‐blockers drugs as compared to placebo or no treatment during the first 24 hours of a cardiovascular event (WMD ‐3.35, 95%CI [‐5.43,‐1.28], p=0.002; I²=60%). The statistically significant difference persisted when a random effects model was applied Figure 15.

15.

Forest plot of comparison: 2 BB, outcome: 2.5 Weighted mean change in Diastolic Blood pressure during first 24 hours.

Heart rate change: There were 5 trials (N=594) included in the heart rate change meta‐analysis. The pooled effect showed a statistically significant greater reduction in HR with beta‐blockers drugs as compared to placebo or no treatment during the first 24 hours of a cardiovascular event (WMD ‐9.68, 95%CI [‐11.99,‐7.37], p<0.00001; I²=4%) Figure 16.

16.

Forest plot of comparison: 2 BB, outcome: 2.6 Weighted mean change in heart rate during first 24 hours.

Calcium channel blocker (CCB)

Systolic blood pressure change: There were 7 trials (N=755) included in the systolic meta‐analysis. The pooled effect showed statistically significant difference in SBP between calcium channel blocker drugs and placebo or no treatment during the first 24 hours of a cardiovascular event (WMD ‐5.49, 95%CI [‐8.42,‐2.56];p=0.0002; I²=80%). The effect was no longer statistically significant when a random effects model was applied.

Diastolic blood pressure change: There were 6 trials (N=565) included in the diastolic meta‐analysis. The pooled effect showed a statistically significant greater reduction in DBP with calcium channel blocker drugs as compared to placebo or no treatment during the first 24 hours of a cardiovascular event (WMD ‐5.08, 95%CI [‐7.00,‐3.15], p<0.0001; I²=64%). The statistical significant difference between groups remained when a random effects model was applied.

Heart rate change: There were 5 trials (N=410) included in the heart rate change meta‐analysis. There was no significant difference in heart rate change between calcium channel blocker drugs and placebo or no treatment during the first 24 hours of a cardiovascular event (WMD ‐1.10, 95%CI [‐4.60,2.40];p=0.54; I²=54%). Figure 17.

17.

Forest plot of comparison: 3 CCB, outcome: 3.6 Weighted mean change in heart rate during first 24 hours.

Nitrates

Systolic blood pressure change: There were 7 trials (N=1,958) included in the systolic meta‐analysis. The pooled effect showed a statistically significant greater reduction in SBP with nitrates drugs compared to placebo or no treatment during the first 24 hours of an acute myocardial infarction (WMD ‐12.67, 95%CI [‐14.51, ‐10.83], p<0.00001; I²=81%). The statistical significance remained when random effects model was applied Figure 18.

18.

Forest plot of comparison: 4 Nitrates, outcome: 4.4 Weighted mean change in Systolic Blood pressure during first 24 hours.

Diastolic blood pressure change: There were 6 trials (N=1146) included in the diastolic meta‐analysis. The pooled effect showed a statistically significant greater reduction in DBP with nitrates drugs as compared to placebo or no treatment during the first 24 hours of an an acute myocardial infarction (WMD ‐7.50, 95%CI [‐9.07,‐5.93], p<0.0001; I²=75%). The statistical significant effect remained when a random effect model was applied .

Heart rate change: There were 6 trials (N=810) included in the heart rate change meta‐analysis. There was no significant difference in heart rate change between nitrates and placebo or no treatment during the first 24 hours of a cardiovascular event (WMD ‐0.83, 95%CI [‐2.83,1.17], p=0.42; I²=57%).

Discussion

Summary of main results

Randomized controlled trials (RCT) comparing an intervention with no treatment or placebo is the method of choice for determining efficacy and effectiveness. The optimal measure of net health effectiveness is the total number of patients with at least one SAE (Wright 2002). However in acute life‐threatening cardiovascular events (CVEs) where the patient is hospitalized, total SAEs are particularly difficult to document and thus rarely measure it. In this review this was proven to be true as no trial reported this outcome. One possible reason is the subjectivity when physician must make a judgment as to whether a particular event led to prolongation of hospitalization as compared to the underlying life‐threatening disease. Therefore, in this acute clinical setting total all‐cause mortality has been considered the best measure of net health effect (Cook 2008). All‐cause mortality is an outcome measure that is not subject to physician judgment and is usually reported in trials. That is the reason that we have focused primarily on total all‐cause mortality in this systematic review.

The primary objective of this review was whether 24 to 48 hour (immediate) administration of blood pressure lowering drugs used within the first 24 hours of an acute cardiovascular event reduced mortality at three different time periods, 2 days, 10 days and ≥30 days. Our initial plan to pool the effects of all blood pressure lowering drugs proved to be inappropriate as the effects of the different classes of drugs was clearly heterogeneous. We have therefore presented the outcome data separately for each of the drug classes.

In studying the immediate intervention, the 2 day mortality provides a measure of the immediate benefits or harms and the 10 day and ≥30 day time periods provides a measure of delayed beneficial or harmful effects that only become manifest later. The principle here has been amply demonstrated by the fibrinolysis trials where early treatment leading to preservation of myocardium leads to reductions in mortality that are not manifest at 2 days, but which increase over time. For example, in ISIS‐2 1988 where 17,187 patients within 24 hours of the onset of a suspected acute myocardial infarction were randomized to receive immediate treatment with a thrombolytic drug or placebo resulted in having no effect on mortality at 48 hours (RR=1.04, 95%CI[0.89,1.21]), but reduced mortality at 10 days (RR= 0.82, 95%CI[0.74,0.91]) and at 35 days (RR=0.77, 95%CI[0.70,0.84]). Unfortunately, in this blood pressure lowering systematic review the number of trials that reported mortality at 10 days and ≥30 days after immediate treatment was insufficient to ascertain whether there were delayed benefits or harms. However, the data that are available suggests that the benefits at 10 days and ≥30 days is less than at 2 days, if anything. Therefore it does not provide any evidence suggesting any significant delayed benefit of immediate BP lowering.

The most important conclusion of this review comes from the mortality data at 2 days after immediate blood pressure lowering. At 2 days there was a highly significant reduction in mortality for nitrates (RR 0.81, 95%CI [0.74,0.89], p<0.0001) and no significant reduction in mortality for the other classes of drugs. Since the data is incomplete, we cannot be very confident whether the hemodynamic effect among the different drug classes, during the first 24 hours, contributed to mortality differences. However, by indirectly comparing the two classes of drugs where statistically significant effects, compared to control, were seen, in these early hemodynamic changes, this review showed that the weighted mean systolic blood pressure lowering effect for beta‐blockers (‐12.54 mmHg) and for nitrates (‐12.67 mmHg) was very similar, whereas the heart rate change was significantly different between the two classes (‐9.68, ‐0.83 bpm, respectively). Despite this similarities and differences in these early hemodynamic changes, the day‐2 mortality effect was favorable for nitrates (RR 0.81, 95%CI [0.74,0.89]) but not for beta‐blockers (RR 0.95, 95%CI [0.85,1.07]).

Unfortunately, we do not have hemodynamic data for ACE‐inhibitors; and the data for CCB was too heterogeneous to yield statistical significance versus control.

This significant mortality benefit at 2 days for nitrates is based predominantly on the two largest trials (ISIS‐4 1995, GISSI‐3 1994), but the authors of those trials have dismissed the finding, as the effect was no longer significant at 35 days. It is thus worth examining these two trials in more detail. The ISIS‐4 1995 trial (N=58,050) was a factorial‐design trial with three independent interventions: isosorbide mononitrate vs. placebo; captopril vs. placebo and magnesium vs. open control. The nitrate used was control‐released isosorbide starting with 30 mg every 12 hours for the first day and followed by a maintenance dose of 60 mg daily for 28 days. However, during the first few days, all patients were allowed to receive non‐study intravenous (IV) nitrates (approximately 47 % of patients in each group ultimately received these). The other large study (N=19,394) was an open‐label, factorial design trial, GISSI‐3 1994, with two independent comparisons: lisinopril vs. control; and glyceryl trinitrate vs. control. In the nitrate group patients were initiated with IV infusion of nitroglycerin at 5 to 20 μg/min (to achieve at least 10% of SBP reduction) during the first 24 hours. After that, the IV infusion was replaced by a transdermal patch providing 10 mg of nitrate per day for 6 weeks. Also in this trial patients were allowed to receive non‐study nitrates. The percentage of patients who received these in the control group was 57.1%, but the corresponding percentage for those allocated to the nitrate group was not reported.

In general, these two trials (ISIS‐4 1995 and GISSI‐3 1994) were similar in terms of population studied but dissimilar in design and drugs used. Despite the differences, the benefit for the nitrates at 2 days, was the same for both trials (RR 0.82, 95%CI [0.73,0.92]) and (RR 0.82, 95%CI [0.68,0.99]), respectively. Furthermore, the fact that the control group also received non‐study nitrates makes it likely that this significant mortality benefit is underestimated . Based on this RCT evidence it is concluded that nitrates should be used routinely within 24 hours of the onset of an acute myocardial infarction as per baseline characteristics outlined in the RCTs.

Unfortunately, since treatment in these trials continued for more than 2 days, these two trials cannot be used to measure the delayed effect of the immediate use of nitrates on mortality at 10 days and ≥ 30 days. In the available trials the immediate use of nitrates had no significant delayed mortality reduction at 10 days, (RR 0.84, 95%CI[0.69,1.01],p=0.07) or ≥30 days(RR 0.92, 95%CI[0.82,1.04],p=0.20) . The lack of a statistically significant effect is partly due to less data, but also due to the fact that the effect estimate appears to be diminishing over time. Thus, this is opposite to the effect seen with fibrinolytic agents and it is some evidence against a delayed benefit of immediate nitrate administration.

On the other hand these large trials, because they continued nitrates for 35 days, can be used to assess the mortality benefit of nitrates given from 3 to 10 days and from 10 to ≥30 days. The effect of nitrates administered from 3 to 10 days is shown in the sensitivity analysis, Figure 19, RR 0.98 [0.91 to 1.06]. This provides evidence for no benefit or harm, when nitrates are administered during this time period after an acute myocardial infarction. Furthermore when the effect of nitrates on mortality was assessed for the period from 10 to ≥30 days it is clear that nitrates are not beneficial (RR 1.10, 95%CI [1.00 to 1.22]). This provides relatively robust evidence for the lack of benefit from nitrates administered beyond day 2 after an acute myocardial infarction. This does not preclude the use of nitrates if other appropriate indications exists such as angina post‐infarction.

19.

Forest plot of comparison: 4 Nitrates, outcome: 4.8 Sensitivity analysis mortality 0‐2 vs. 3‐10 days vs. day 11 to end of treatment (35‐42 days).

The present systematic review did not find statistical evidence that immediate treatment with ACE inhibitors reduces mortality at 2 days after a myocardial infarction. Although, most of the evidence comes from the same 2 large trials where the nitrates were studied, the evidence is of borderline significance (RR 0.91, 95%CI [0.82,1.00], p=0.05), so that it remains possible that a real benefit has been missed. Other possible explanation is that ACEi could increase other serious adverse events in this early period. Evidence in favor of this possibility in ISIS‐4 1995 is that cardiogenic shock or profound hypotension requiring termination of study treatment was significantly higher in the ACE‐inhibitor group as compared to placebo (p<0.01 and p<0.001, respectively) during the first 24‐48 hours. It is also possible that the beneficial effect from ACE‐inhibitors is specific to certain sub‐groups of patients such as those patients with anterior infarction (ISIS‐4 1995). If that was the case the significant beneficial effect would be diluted in the overall effect. The best way to resolve this uncertainty is to do an individual patient meta‐analysis of the available trials to try to identify subgroups of patients who benefit from routine use of ACE‐inhibitors in this early phase of an acute MI.

Unfortunately, there were only two small trials in which a delayed effect of the immediate treatment with ACE‐inhibitors could be assessed. Therefore, there was insufficient RCT evidence to ascertain the delayed effects of these drugs when given as immediate treatment (within 24 hours and lasting for up to 48 hours) following an acute myocardial infarction.

The present systematic review demonstrated that the immediate treatment with BB was not associated with significant mortality benefit at 2 days (RR 0.95, 95%CI [0.85, 1.07], p=0.39), 10 days (RR 1.12, 95%CI [0.60, 2.07], p=0.73) or ≥30 days. However, there was very little data to assess the delayed effect of immediate treatment with beta‐blockers. In addition there was significant heterogeneity for the day‐2 mortality outcome (I²=67%). The heterogeneity was not explained by the type of beta‐blocker. However, when a pre‐planned sensitivity analysis was performed according to trial design (double‐blind; open‐label), BB were not associated with a significant reduction in mortality as compared with placebo or no treatment among 51,814 patients in double‐blind trials (RR 1.04, 95%CI [0.91,1.19]; I²=30%) and was associated with a significant reduction in mortality among 16,193 patients in open‐label trials (RR 0.73, 99%CI [0.58,0.91]; p=0.006; I²=69%). Given the clear lack of benefit in the double‐blind trials, the most likely explanation for the benefit of beta blockers in the open label trials is that the two groups had other treatment differences than the planned intervention (performance bias). In keeping with that possibility, significantly more patients in the control group as compared to the beta‐blocker group received calcium channel blockers drugs (RR 1.89 [1.74,2.06]; p<0.00001) in ISIS‐1 1986. If calcium channel blockers increase mortality this would lead to false evidence of a mortality benefit for beta‐blockers. The evidence of a clear lack of mortality benefit in the more robust double‐blind trials makes us confident that routine use of beta‐blockers early after an MI is not useful.

The present systematic review also showed that immediate treatment with CCBs was not associated with a statistically significant reduction in all‐cause mortality at 2 days, 10 or ≥30 days. This conclusion is weak due to the fact that the included trials do not report mortality at all of these times and because there were many CCB trials excluded from this meta‐analysis, due to lack of reporting or improperly reported mortality data. The reason for this lack of or improper reporting of mortality is not known, however, we feel it is unlikely that trials demonstrating a mortality benefit would not have reported their results in full. We suspect that publication bias exists as we excluded 84% of the calcium channel blocker trials due to a lack of or insufficient mortality reporting. Therefore, if anything, we think that the data showing no benefit or harm from CCBs in this review is likely biased and that the full mortality data would show a statistically significant harm from CCBs.

The effects of short‐term treatment provide additional information to what we have learned from the immediate treatment. Nitrates were again associated with a significant reduction in mortality at 10 days, however, as discussed above, this reduction was entirely accounted for by the immediate treatment reduction in mortality at 2 days. In contrast, ACE inhibitors after short‐term treatment now show a significant effect at 10 days (RR 0.93, 95%CI [0.87,0.98]). The reason for the significant mortality benefit at 10 days is that, in contrast to the nitrates, in the two large trials (GISSI‐3 1994; ISIS‐4 1995), the effect of the continued administration of ACE inhibitors was similar from 0 to 2 days (RR 0.91 ), from 3 to 10 days (RR 0.93, 95%CI [0.86 to 1.01]); and from day 11 to day 35 or 42 (RR 0.93, 95%CI [0.85 to 1.03]), see Figure 20. This resulted in a significant mortality benefit for ACE inhibitors continued for 35‐42 days (RR 0.93, 95%CI [0.88 to 0.97]) as compared to a non‐significant mortality effect of nitrates continued for 35‐42 days (RR 0.97, 95%CI [0.92 to 1.02]). However, it is emphasized that in a trial where it was possible to compare nitrates vs. ACE inhibitors in a head to head comparisons, GISSI‐3 1994, there was no differences in mortality at 42 days (7% vs. 6.6%). This led us to conclude that the routine use of ACE inhibitors post MI reduces mortality to a modest degree. However, it remains uncertain as to the optimal time of starting these drugs, and whether ACE inhibitors should be targeted at a subgroup of patients where the mortality benefit is larger.

20.

Forest plot of comparison: 1 ACEi, outcome: 1.7 Sensitivity analysis: mortality 0‐2 vs. 3‐10 days vs day 11 to end of treatment (35‐42 days).

The data to assess the effects of short‐term treatment (lasting maximum 10 days) on mortality at ≥ 30 days was sparse for nitrates (7 trials) and nil for the ACE‐inhibitors.

In the present systematic review, short‐term treatment with beta‐blockers showed no statistically significant effect on mortality at 10 days (RR 0.96, 95%CI[0.91,1.02]; 14 trials, N=71,457), but a statistically significant effect at ≥ 30 days (RR 0.91,95%CI [0.84,0.99], p=0.03; 5 trials, N=18,373). Although the effect at ≥ 30 days is statistically significant, suggesting a delayed benefit with beta‐blockers, we believe this is a chance effect and probably not real.  When mortality for these five trials was analyzed separately for the periods 0‐10 days and 11 to ≥ 30 days, a significant benefit is seen only during the first 10 days, contradicting the "suggested delayed benefit" shown above. Furthermore these results contradict the lack of significant mortality effect during the first 10 days, when all 14 trials with available data are considered. A potential explanation of these contradictions is publication bias from the 5 trials that reported data at ≥ 30 days. In addition, the apparent exaggerated benefit in two trials (ISIS‐1 1986 and Yusuf 1983) is possible due to performance bias as in both trials significantly more patients received calcium channel blockers (CCBs) in the control group. This could potentially increase mortality in the control group and exaggerate the benefit from the beta‐blocker. The adverse effect of CCBs is in accordance with the findings of this review (see below).

Our systematic review showed that short‐term treatment with CCBs was not associated with a statistically significant reduction in all‐cause mortality at 10 or ≥30 days. Similar to the effect of immediate treatment, this conclusion is not robust because only a small percentage of CCB trials could be included in this review. In our sensitivity analyses, there was a trend towards a greater mortality among MI patients treated short‐term with calcium channel blockers as compared to placebo (RR 1.60 95%CI[0.90, 2.86]); particularly in those receiving dihydropyridine CCBs (RR 1.91, 95%CI[0.98, 3.72]).

Only 5 acute stroke RCTs were included in the present systematic review and all of these studied calcium channel blocker (CCB) drugs. CCBs were not associated with significant effect on mortality at 10 days (RR 0.81, 95%CI [0.54,1.21]) in this setting. This evidence is clearly insufficient to know whether CCBs are beneficial or harmful in acute stroke patients.

Overall completeness and applicability of evidence

Out of 135 potentially appropriate randomized controlled trials we excluded 65 studies in which mortality data was non‐extractable or usable for our review. Thus this data, if made available by the authors, could be added to our review . Of these 65 excluded trials, 14 involved an ACE inhibitor (N=4,407); 21 trials involved a B‐adrenergic Blocker (N=2,524); 8 trials involved a nitrate (N=1,475) and 20 trials involved a CCB (N=10,958). These excluded patients make up a small proportion of the available included patients, for the ACE inhibitors, 5.0% ; BBs 3.4% ; and nitrates 1.7% , but a large proportion of the calcium channel blocker population, 84% . Due to the relatively small number of excluded patients and trials for the ACEi, BB and nitrates groups, we are pretty confident that the effect sizes calculated in these meta‐analysis are a good estimate of the true effect. However, that is not the case for CCBs, where it is more likely that the non‐significant increase in short‐term mortality among those patients with acute myocardial infarction found in this review (RR 1.60 95%CI[0.90, 2.86]) would become statistically significant.

We have no information about the effects on mortality of other blood pressure lowering drugs (such as clonidine, prazosin, hydralazine, etc) used within 24 hours of the onset of an acute myocardial infarction or stroke. Likewise, we have no information regarding the effects on mortality produced by drugs embraced in this review started within 24 hours of the onset of other cardiovascular events such as acute aortic dissection, acute pulmonary edema, sub‐arachnoid hemorrhage or unstable angina because mortality data was not assessed or reported in these conditions. To the best of our knowledge there is only one randomized controlled trial including only patients with acute aortic dissection and randomizing the patients within 24 hour of the onset (Yoshida 1998). However, this trial compared two active treatments. 
 
 Thus, this systematic review evidence is primarily applicable to patients with acute myocardial infarction receiving blood pressure lowering drugs within 24 hours of the onset of the event and having a level of risk similar to those participants from the randomized trials included in this review. Since the evidence for effectiveness of nitrates and ACE‐inhibitors mostly comes from two large trials (ISIS‐4 1995; GISSI‐3 1994) the practical recommendations primarily apply to patients having the following characteristics:

ISIS‐4 1995:

Patients with suggestive of suspected or definite acute MI (with or without electrocardiographic changes); within 24 hours of symptom onset; and with no clear indications for, or clear contraindications to any of the trial treatments.

In this trial the contraindications were not set by the protocol, but by the responsible physician and might include: negligibly low risk of MI death (e.g. normal electrocardiogram), major life‐threatening disease other than acute MI, or a high risk of adverse effects of trial treatment such as: decreased blood pressure‐ cardiogenic shock or severe hypotension (eg., SBP persistently <90‐100 mm Hg, especially with right ventricular infarction or poor peripheral perfusion); severely decreased plasma volume‐clinical evidence of severe fluid depletion, perhaps due to chronic diuretic use.

Overall 74% were male, 28% were 70 or older, 92% were ultimately confirmed to have an AMI, 86% had Killip class I (no signs or symptoms of heart failure), and 17% had previous MI.

The co‐interventions given in the overall trial were: non‐study intravenous nitrates (˜47%), anti‐platelets (˜93%), fibrinolytic agents (˜68%), intravenous beta‐blockers (˜9%).

GISSI‐3 1994:

Inclusion criteria: patients with chest pain accompanied by ST elevation or depression of at least 1 mm in one or more peripheral leads of the electrocardiogram (ECG), or at least 2 mm in one or more precordial leads; within 24 hours of symptoms onset; and having no contraindications to the study treatments. Exclusion criteria: severe heart failure requiring any of the study treatments; Killip class 4; high risk of further serious haemodynamic deterioration after treatment with vasodilators (SBP ≤ 100 mm Hg), contraindications to study drugs ‐namely a history of clinically relevant renal failure (serum creatinine ≥ 177 mol/L, proteinuria > 500 mg per 24 hours or both), history of bilateral stenosis of the renal arteries, allergies to one of the study drugs, other life threatening disorders (eg. tumors, serious respiratory diseases).

Overall 78% were male, 27% were 70 or older, 93% were ultimately confirmed to have an AMI, 83% had Killip class I, and 14% had previous MI.

The co‐interventions given in the overall trial were: non‐study intravenous nitrates (˜57%‐ only reported for the control group), aspirin (˜84%), fibrinolytic agents (˜72%), intravenous beta‐blockers (˜30%)

Quality of the evidence

Forty trials (60%) out of 65 included studies were double‐blind, involving 75% (N=125,487) of the entire studied population. Although, the number of trials reporting adequate concealment of allocation was relatively small [13/40 (32%) in double‐blind and 10/25 (40%) in open‐label design trials] the number of patients randomized with concealment of allocation was very high (121,618 in double‐blind trials out of a total 125,487 and 37,055 in open‐label trials out of a total 40,719 ). Similarly, although more than 75% of trials had an incomplete or not reported all‐cause mortality at 2 days, the studies that indeed reported it were large enough (N=131,603) to decrease this risk of incomplete outcome to a minimum. That was not the case for ≥30 day mortality which was reported by less than 25 % of trials and the trials were not large enough (24,918) to dismiss the possibility of bias for this outcome.

Potential biases in the review process

One limitation of the present review is that we did not consider many trials, which claimed enrolling their patients in the early period of an acute cardiovascular event, i.e. within 48 to 72 hours. Evidently in these types of studies it is anticipated that some patients could be enrolled and have started their study treatment within 24 hours; therefore, data from these patients are missing from our review. Including such patients would require obtaining individual patient data from these reviews, which was not possible. We have tried to get information from some recent trials of this type (for example from the ACCESS 2003 trial), but with no success. Even though we accept this as a limitation, we feel that adding these trials would create a more unacceptable limitation as we are interested in determining the specific effects of early treatment and we do not want to contaminate that with patients enrolled after the early vulnerable period.

Another limitation of this review is that we assessed blood pressure and heart rate changes only for the first 24 hours of the initiation of treatment. The objective was to assess whether these early changes can be related to mortality at 2, 10 or ≥30 days. Expanding the time window for these measures would likely have made more data available, however we think that the effects during the first 24 hours are the most important. An additional problem was encountered in these acute settings where significant number of patients are censored in terms of BP and heart rate data due to death, withdrawals or losses to follow‐up. Therefore it is difficult to have BP data in the same population that is contributing to the mortality data. We are aware that due to the fact that BP and heart rate data were only available from certain number of trials we do not have a high degree of confidence in the magnitude of these measures.

Agreements and disagreements with other studies or reviews

Nitrates:

There is a previous overview and meta‐analysis that assessed the effects of nitrates in patients with acute myocardial infarction (Yusuf 1988). With only 2,000 patients they found a significant effect (OR 0.55, 95%CI [0.39,0.76]) on early (first week or in hospital) mortality. These authors did not distinguish between immediate treatment or short‐term treatment. Our overall results at 10 days (or hospital) including both immediate and short‐term treatment is not in accordance with their effect estimate: OR 0.90,95%CI [0.85,0.96]p=0.0008. Our results with the inclusion of more recent larger trials in 84,185 patients is clearly a better estimate of the true treatment effect.

Our results are also in disagreement with the guidelines from the the American College of Cardiology and the American Heart Association (ACC/AHA, 2004). In chapter VI regarding the initial management of MI in the emergency department these guidelines have relegated the use of nitrates (nitroglycerin exclusively) to patients with ongoing ischemic discomfort or control of hypertension or management of pulmonary congestion.

The present systematic review demonstrates that the use of nitrates within 24 hours significantly reduces all‐cause mortality (RR 0.81, 95%CI [0.74,0.89], p<0.0001) at 2 days. This is consistent with 4 or 8 deaths prevented for 1000 low or high risk patients treated, respectively (see Table 1). Therefore, routine nitrates should be administered to all patients with suspected acute myocardial infarction, who do not have the specific contraindications outlined above. Due to its proven effectiveness in the ISIS‐4 1995 trial and ease of administration, controlled‐release oral isosorbide‐5‐mononitrate 30 mg twice a day the first day and 60 mg the second day is the best choice. However, the use of routine nitrates should be restricted to the first 48 hours as there was no further reduction in mortality associated with their use beyond day 2.

The mechanism whereby nitrates reduce mortality in the immediate period after an acute MI and not after that is not known. Since it appears from this review that blood pressure reduction is not the explanation, there must be another mechanism. Nitrates have been used for more than a century to relieve myocardial ischemia and chest pain. This has been thought to be due to their vasodilatory effects: venous, arterial, coronary vessels and on the redistribution of blood flow towards the ischemic subendocardium. These mechanisms may explain the early mortality benefit.

A second possible mechanism is that nitric oxide donors such as nitrates and nitroprusside inhibit platelet adhesion and aggregation. In the late 1980's in vivo animal experiments demonstrated that an infusion of nitroglycerin reduced platelet deposition on damaged arteries (Lam 1988). These findings were confirmed in the early 1990's in vivo (De Caterina 1990) showing that isosorbide mono‐nitrates produced a dose‐dependent inhibition of platelet aggregation and thromboxane production induced by adenosine diphosphate and adrenaline in patients with objectively proven coronary artery disease. A controlled study (Butterworth 1998) demonstrated that nitroprusside given at a dose which reduced MAP by 10 mmHg, significantly inhibited platelet aggregation and improved regional cerebral blood flow. Furthermore, the results from ISIS‐4 1995 and ESPRIM 1994 trials support this theory, as they showed a greater mortality benefit with nitrates in patients who never received anti‐platelets drugs. For example, in the former, the difference in mortality between the nitrate group and placebo group in patients who never received anti‐platelets was 15.7% vs. 17.7%, respectively (RR 0.89). In contrast, this difference was negligible in patients who did receive anti‐platelet drugs, 6.9% vs. 7.0%, respectively (RR 0.99). An explanation for the early and not late nitrate benefit is that the activation of the sympathetic nervous system (outflow of catecholamines, epinephrine and norepinephrine) is increased in the immediate period of myocardial infarction (Karlsberg 1981). At this early time catecholamines would be the most prominent inductive mechanism for platelet aggregation. It has been shown that nitric oxide donors such as nitrates exert a platelet aggregation inhibitory response only induced by ADP and catecholamines (De Caterina 1990). That is, nitrates do not interfere with the ongoing thrombotic stimulus induced by thrombin, collagen or other stronger inducers of thromboxane generation and platelet aggregation.

A third potential explanation for the early mortality benefit conferred by nitrates is that, in addition to their hemodynamic and cardiac work‐load benefits, they might interfere with the heart excitability and/or conduction system in the early hours following the myocardial infarction. Before the identification of the endothelium‐derived relaxing factor as nitric oxide in the late 80's there were some reports showing that nitrates have anti‐arrhythmic effects in acute myocardial ischemia animal models( Borer 1974; Stockman 1979; Cano 1986). At that time the theory that nitrates work through the release of nitric oxide (NO) increasing guanosine 3':5'‐ cyclic monophosphate (cGMP) was presented (Katsuki 1977; Kukovetz 1979). More recently, others have reported the effects of nitric oxide on the excitability of the autonomic cells (sinoatrial node or atrioventricular node Exner 1999; Han 1994; Martynyuk 1996; Han 1997; Musialek 1997) or directly on myocytes (Brahmajothi 2007). The fact that neural nitric oxide synthase (nNOS) has now been identified in the nerve fibers of the heart (Tanaka 1998, Hassall 1992), suggests a broader role of NO in the heart's electrophysiology . An augmented expression on nNOS during acute myocardial infarction in animal models has been demonstrated (Takimoto 2002). Thus, an anti‐arrhythmic effect from nitrates/nitric oxide remains a possibility (Brack 2007; Chowdhary 2004; Fei 1997; Pabla 1996).

ACE‐inhibitors

The effects of ACEi on AMI patients have been reported in two systematic reviews (AMICG 1998 ; Rodrigues 2003). In the former, 98,496 patients were included from trials comparing ACEi vs. placebo or no treatment. The authors demonstrated a mortality reduction with ACEi obtained at 7 days and 30 days (RR 0.92, 95%CI [0.86,0.97]; and RR 0.93, 95%CI [0.89,0.98], respectively. Thus, these 10 day findings are very similar to ours. The other ACEi systematic review (Rodrigues 2003) had the objective to assess the effect of long‐term treatments as they did not limited the length of treatment in the trials and quantified mortality at 30 days, 6 months and 1 year. Thus, this review cannot be compared to ours.

The American College of Cardiology and the American Heart Association (ACC/AHA, 2004) have recommended that an ACEi should be administered orally within the first 24 hours of STEMI (ST‐elevation myocardial infarction) to patients with anterior infarction, pulmonary congestion or LVEF less than 0.40 in the absence of hypotension or known contraindications to that class of medications. However, in a second paragraph they further state that ACEi administered within the first 24 hours of STEMI can be useful without the above characteristics.

In our ACEi analysis the effect estimates for open‐label trials and double‐blind trials differ at 2 days RR 0.86, 95%CI [0.71,1.03]) and RR 0.93, 95%CI [0.83,1.04]), respectively, calling into question the borderline overall mortality benefit at 2 days (RR 0.91, 95%CI [0.82,1.00]), p=0.05. Furthermore, nitrates confer a superior relative 0.81 vs. 0.91, and absolute mortality benefit 4‐8 per 1000 vs. 2‐4 per 1000, respectively. Thus, we feel that the optimal time of administration of ACE inhibitors post acute MI is presently unknown and further research is required.

Beta‐adrenergic antagonists (BB)

Other systematic reviews (SRs) have evaluated the effect of beta‐blockers (BB) in AMI (Yusuf 1985, Freemantle 1999, Al‐Reesi 2008). All of these reviews have accepted trials where treatment was started later than 24 hours after the onset of AMI. Yusuf 1985 concluded that reliable estimation of the effects of early beta blockade on mortality has not yet been achieved. Freemantle 1999 also concluded that there were no mortality benefits from BB in "short‐term" trials. The most recent review (Al‐Reesi 2008), inexplicably excluded 11 trials that we included in our review. Despite that, their conclusion stating that "acute intervention with B‐blockers does not result in statistical significant short‐term survival benefit following AMI" is in agreement with ours.

Despite these systematic reviews ACC/AHA, 2004 recommended that oral beta‐blocker therapy should be administered promptly to those suspected MI patients without a contraindication irrespective of concomitant fibrinolytic therapy or performance of primary PCI.

Based on our systematic review we do not recommend routine BB for the immediate or short‐term treatment of patients with suspected MI. However, it should be emphasized that these findings and recommendations are not contradictory to the long‐term mortality benefits for BB post MI when these drugs are started a few days or weeks after myocardial infarction and continued for some months (Yusuf 1985 and Freemantle 1999) .

Calcium Channel Blockers (CCB)

In Acute Myocardial infarction

The one other systematic review that has assessed CCB drug in patients with acute myocardial infarction or unstable angina (Held 1989) also included trials in which treatment was started after 24 hours or trials where the time of entry was not specified. However, they also concluded that "CCBs do not reduce the risk of initial or recurrent infarction or death when given routinely to patients with acute myocardial infarction or unstable angina".

ACC/AHA, 2004 have restricted the use of CCB by stating that it is reasonable to give verapamil or diltiazem to patients in whom beta‐blockers are ineffective or contraindicated for relief of ongoing ischemia or control of a rapid ventricular response with atrial fibrillation of flutter after STEMI in the absence of CHF, LV dysfunction or atrioventricular (AV) block.

In Stroke

There is one Cochrane systematic review already published that assessed CCBs in acute stroke (Horn 2000) but with a different research question from ours. Their focus was not limited to the immediate initiation of treatment and for short term treatment. They included 28 studies. Of those, 12 trials had required their patients to start treatment within 24 hours of the onset. Of those, we have excluded one trial because it compared IV infusion vs. oral treatment, rather than CCB vs. placebo; and 5 trials because mortality data was not reported or usable in our review. Despite the differences their conclusion "No evidence is available to justify the use of calcium antagonists in patients with acute ischemic stroke" agrees with ours.

Two Cochrane systematic reviews involving patients with acute stroke and blood pressure lowering drugs have been published (Bath 2002; Geeganage 2008). The former assessed nitrates for acute stroke with different methodology and objectives than ours. Studies were not limited to truly randomized trials or to the immediate initiation of treatment and for short term treatment. Since only 2 studies were included the authors concluded that there was insufficient evidence to recommend the use of nitrates. We did not include those two trials because treatment started days after the onset of the stroke. Geeganage 2008 also had a different approach to ours, as it was not limited to RCTs studying blood pressure lowering drugs. In addition, in this review RCTs are not limited to a certain time of starting treatment after the stroke (for example, it included trials where treatment started even 7 days after the onset).  Out of their 12 included studies, 3 trials started the treatment within 24 hours. We excluded all these 3 for the following reasons: in the INTERACT 2008 trial, patients were not allocated to a class of drug vs. placebo or no treatment. Instead, patients were allocated to two different BP targets (140 vs. 180 mmHg) to be achieved in 1 hour and maintained for 7 days. The ACCESS 2003 trial included patients within 24 and 36 hours but without separating their results according to these times.  And in the third trial (Eveson 2007) results were not based on ITT principles. Patients first were randomized and then withdrawn if the diagnosis was wrong. Analysis was not performed based on all randomized patients.

Authors' conclusions

Implications for practice.

Acute myocardial infarction

• Nitrates administered within 24 hours of symptom onset significantly decrease day 2 all‐cause mortality (4 to 8 deaths prevented per 1000).

• The evidence shows that continuation of nitrates beyond day 2 does not reduce mortality.

• ACE inhibitors administered within 24 hours of symptom onset have not been shown to significantly reduce mortality at 2 days.

• ACE inhibitors administered within 24 hours of symptom onset and continued for 10 days significantly reduce day 10 all‐cause mortality (3 to 5 deaths prevented per 1000).

• The optimal time of starting ACE inhibitor therapy post myocardial infarction is not known.

• Beta‐blockers started within 24 hours of symptom onset do not reduce all‐cause mortality at 2 days, or after short‐term use at 10 days .

• Calcium channel blockers started within 24 hours of symptom onset do not decrease mortality and a trend towards increased mortality was seen after short‐term use of these drugs at 10 days (RR 1.60 95%CI[0.90, 2.86]).

Other acute cardiovascular conditions

• Calcium channel blockers administered after acute stroke have not been shown to affect mortality but the data is insufficient.

• There is no RCT information regarding the effects on mortality produced by blood pressure lowering drugs started within 24 hours of the onset of other cardiovascular events such as acute aortic dissection, acute pulmonary edema, unstable angina, intracranial or sub‐arachnoid hemorrhage.

Implications for research.

In patients with acute myocardial infarction:

• Future RCTs of early treatment should report mortality at standard times: 2 days, 10 days, 30 days and 6 months.

• Mortality data at the above times from all RCTs should be made available. This is particularly important for the existing calcium channel blocker RCTs.

• Future trials in this condition need to be cognizant of the possibility that treatment effects may be different during the first 2 days as compared with after 2 days as demonstrated in this review.

• An individual patient meta‐analysis of immediate treatment ACE inhibitor trials is needed to ascertain whether there is a subgroup of immediate post MI patients with a mortality benefit.

• More RCTs are needed to better define the optimal time to start ACE inhibitor therapy.

In patients with acute stroke, unstable angina, acute pulmonary edema, cerebral hemorrhage or acute aortic dissection:

• Because blood pressure lowering drugs are frequently used in these settings there is a need for more large RCTs assessing different aspects of these interventions: e.g. different drug classes, different drugs within a class, dosing regimens, timing of onset of treatment, blood pressure threshold for treatment, etc.

• Regardless of the design of these RCTs, such trials must report total all‐cause mortality at standard different times of follow‐up: 2 days, 10 days, 1 month and 6 months.

• An international organization should standardize and mandate the documentation and reporting of total serious adverse events in hospitalized and critically‐ill patients.

What's new

Date	Event	Description
26 January 2010	Amended	1. Corrections in the conclusions; Re: ACE‐inhibitors: Instead of reading 2‐4, it should read 3‐5. Re: CCBs: Instead of reading RR 1.57, it should read 1.60 (CI also changed). 2. Correction in the agreements and disagreements with other studies section; Re: Nitrates: wording in the 5th paragraph 3. Correction in effect of interventions according to outcomes section; Re: Beta‐Blockers, mortality at ≥30 days Instead of reading (RR 0.92,95%CI [0.85,1.00], p=0.05, I2=0%), it should read (RR 0.91,95%CI [0.84,0.99], p=0.03, I2=0%). Accordingly to the above, the third last paragraph of the summary of main results was modified. 4. Correction in the Mortality reported for Yusuf et al 1983. 5. Correction in the Mortality reported for Durrer et al 1982.

History

Protocol first published: Issue 3, 2007
 Review first published: Issue 4, 2009

Date	Event	Description
12 June 2008	Amended	Converted to new review format.

Appendices

Appendix 1. Search Strategies

MEDLINE SEARCH

controlled clinical trial.pt.

randomized.ab.

drug therapy.fs.

randomly.ab.

trial.ab.

groups.ab.

allocat$.ab.

or/1‐7

animals.sh.

8 not 9

Alacepril.mp.

Benazepril.mp.

captopril.mp.

ceronapril.mp.

cilazapril.mp.

derapril.mp.

enalapril.mp.

enalaprilat.mp.

fosinopril.mp.

idapril.mp.

imidapril.mp.

Lisinopril.mp.

moexipril.mp.

moveltopril.mp.

perindopril.mp.

quinapril.mp.

ramipril.mp.

spirapril.mp.

temocapril.mp.

trandolapril.mp.

zofenopril.mp.

angiotensin converting enzyme inhibitor.mp. or Angiotensin‐Converting Enzyme Inhibitors/

acebutolol.mp.

atenolol.mp.

Bisoprolol.mp.

esmolol.mp.

labetalol.mp.

metoprolol.mp.

nadolol.mp.

practolol.mp.

propranolol.mp.

sotalol.mp.

timolol.mp.

carvedilol.mp.

Adrenergic beta‐Antagonists.mp.

Amlodipine.mp.

Aranidipine.mp.

Azelnidipine.mp.

Barnidipine.mp.

Bencyclane.mp.

Benidipine.mp.

Bepridil.mp.

Cilnidipine.mp.

Cinnarizine.mp.

Clentiazem.mp.

Darodipine.mp.

Diltiazem.mp.

Efonidipine.mp.

Elgodipine.mp.

Etafenone.mp.

Fantofarone.mp.

Felodipine.mp.

Fendiline.mp.

Flunarizine.mp.

Gallopamil.mp.

Isradipine.mp.

Lacidipine.mp.

Lercanidipine.mp.

Lidoflazine.mp.

Lomerizine.mp.

Manidipine.mp.

Mibefradil.mp.

Nicardipine.mp.

Nifedipine.mp.

Niguldipine.mp.

Nilvadipine.mp.

Nimodipine.mp.

Nisoldipine.mp.

Nitrendipine.mp.

Perhexiline.mp.

Prenylamine.mp.

Semotiadil.mp.

Terodiline.mp.

Tiapamil.mp.

verapamil.mp.

calcium channel blocker.mp. or Calcium Channel Blockers/

nitroprusside.mp.

nitroglycerine.mp.

Nitroglycerin/ or nitroglycerine.mp. or Isosorbide Dinitrate/

nitrates.mp. or Nitrates/

urapidil.mp.

Trimethaphan/ or trimethaphan camsylate.mp.

reserpine.mp.

phentolamine.mp.

methyldopa.mp.

labetalol.mp.

ketanserine.mp.

hydralazine.mp.

guanethidine.mp.

fenoldopam.mp. or FENOLDOPAM/

diazoxide.mp.

clonidine.mp.

thiazide$.mp.

hydrochlorothiazide.mp.

chlorthalidone.mp. or Chlorthalidone/

furosemide.mp. or Furosemide/

or/11‐106

10 and 107

unstable angina.mp.

acute left ventricular failure.mp.

Pulmonary Edema/ or pulmonary oedema.mp.

stroke.mp.

(acute adj2 stroke).ti,ab.

Aneurysm, Dissecting/ or aortic dissection.mp.

Intracranial Hemorrhages/ or Cerebral Hemorrhage/ or intracranial haemorrhage.mp.

Intracranial Aneurysm/ or Subarachnoid Hemorrhage/ or subarachnoid haemorrhage.mp.

acute myocardial infarction.mp.

or/109‐117

myocardial infarction.ti,ab.

acute.ti,ab.

threatened.ti,ab.

suspected.ti,ab.

or/120‐122

119 and 123

118 or 124

108 and 125

placebo$.mp.

(control$ adj2 treatment$).ti,ab.

(treatment$ adj2 control$).ti,ab.

(open adj2 control).ti,ab.

(served adj2 control).ti,ab.

(control adj2 group).ti,ab.

(deferred adj2 group).ti,ab.

(control$ adj2 subject$).ti,ab.

(control$ adj2 patient$).ti,ab.

or/127‐135

126 and 136

EMBASE SEARCH

randomized controlled trial.mp.

randomized controlled trials.mp.

controlled clinical trial.mp.

controlled clinical trials.mp.

randomized.ab.

random$.ab.

allocat$.ab.

((singl$ or doubl$ or trebl$ or tripl$) adj25 (blind$ or mask$)).mp.

trial.ab.

groups.ab.

or/1‐10

exp animal/

11 not 12

Alacepril.mp.

Benazepril.mp.

captopril.mp.

ceronapril.mp.

cilazapril.mp.

derapril.mp.

enalapril.mp.

enalaprilat.mp.

fosinopril.mp.

idapril.mp.

imidapril.mp.

Lisinopril.mp.

moexipril.mp.

moveltopril.mp.

perindopril.mp.

quinapril.mp.

ramipril.mp.

spirapril.mp.

temocapril.mp.

trandolapril.mp.

zofenopril.mp.

angiotensin converting enzyme inhibitor.mp. or Angiotensin‐Converting Enzyme Inhibitors/

acebutolol.mp.

atenolol.mp.

Bisoprolol.mp.

esmolol.mp.

labetalol.mp.

metoprolol.mp.

nadolol.mp.

practolol.mp.

propranolol.mp.

sotalol.mp.

timolol.mp.

carvedilol.mp.

Adrenergic beta‐Antagonists.mp.

Amlodipine.mp.

Aranidipine.mp.

Azelnidipine.mp.

Barnidipine.mp.

Bencyclane.mp.

Benidipine.mp.

Bepridil.mp.

Cilnidipine.mp.

Cinnarizine.mp.

Clentiazem.mp.

Darodipine.mp.

Diltiazem.mp.

Efonidipine.mp.

Elgodipine.mp.

Etafenone.mp.

Fantofarone.mp.

Felodipine.mp.

Fendiline.mp.

Flunarizine.mp.

Gallopamil.mp.

Isradipine.mp.

Lacidipine.mp.

Lercanidipine.mp.

Lidoflazine.mp.

Lomerizine.mp.

Manidipine.mp.

Mibefradil.mp.

Nicardipine.mp.

Nifedipine.mp.

Niguldipine.mp.

Nilvadipine.mp.

Nimodipine.mp.

Nisoldipine.mp.

Nitrendipine.mp.

Perhexiline.mp.

Prenylamine.mp.

Semotiadil.mp.

Terodiline.mp.

Tiapamil.mp.

verapamil.mp.

calcium channel blocker.mp. or Calcium Channel Blockers/

nitroprusside.mp.

nitroglycerine.mp.

Nitroglycerin/ or nitroglycerine.mp. or Isosorbide Dinitrate/

nitrates.mp. or Nitrates/

urapidil.mp.

Trimethaphan/ or trimethaphan camsylate.mp.

reserpine.mp.

phentolamine.mp.

methyldopa.mp.

labetalol.mp.

ketanserine.mp.

hydralazine.mp.

guanethidine.mp.

fenoldopam.mp. or FENOLDOPAM/

diazoxide.mp.

clonidine.mp.

thiazide$.mp.

hydrochlorothiazide.mp.

chlorthalidone.mp. or Chlorthalidone/

furosemide.mp. or Furosemide/

or/14‐109

13 and 110

unstable angina.mp.

acute left ventricular failure.mp.

Pulmonary Edema/ or pulmonary oedema.mp.

stroke.mp.

(acute adj2 stroke).ti,ab.

Aneurysm, Dissecting/ or aortic dissection.mp.

Intracranial Hemorrhages/ or Cerebral Hemorrhage/ or intracranial haemorrhage.mp.

Intracranial Aneurysm/ or Subarachnoid Hemorrhage/ or subarachnoid haemorrhage.mp.

acute myocardial infarction.mp.

or/112‐120

myocardial infarction.ti,ab.

acute.ti,ab.

threatened.ti,ab.

suspected.ti,ab.

or/123‐125

122 and 126

121 or 127

111 and 128

placebo$.mp.

(control$ adj2 treatment$).ti,ab.

(treatment$ adj2 control$).ti,ab.

(open adj2 control).ti,ab.

(served adj2 control).ti,ab.

(control adj2 group).ti,ab.

(deferred adj2 group).ti,ab.

(control$ adj2 subject$).ti,ab.

(control$ adj2 patient$).ti,ab.

or/130‐138

129 and 139

CENTRAL SEARCH

This search was identical to EMBASE and MEDLINE searches but without the "trials" component as CENTRAL only deals with randomized‐related studies

Appendix 2. Replies from trialists

From: Pietro Di Pasquale [mailto:LEHDI@tin.it] 
 Sent: June 28, 2008 7:22 AM 
 To: maiperez@interchange.ubc.ca 
 Subject: by p. di pasquale

Dear Marco

Thank you for your letter.

The first study (int J cardiol 1994;43;43‐50) was performed to   verify the addition of captopril in AMI patients. This is a single blind study and end points were reperfusion damage and clinical outcome.

The other study (Int J Cardiol 1994; 46:107‐112) was an extension of the first study. In fact ,in this study we analyzed  the AMI patients enrolled in first study to control  the effect of beta‐blocker treatment (alone or associated with captopril or placebo). In fact, it was not possible to perform a randomization because in 1994 beta‐blockers in acute phses of AMI  was  reccomnede treatment.

The two studies were reports of the same trial  and studied  different therapeuthic associations.

Pietro Di Pasquale MD.

P.S. In the  2° study the patients with inferior AMI were excluded. This is the reason  for lower number of patients in the 2° study.(84 pts versus 82 pts)

        In the first study the randomized patients were 131 vs 128 (anterior AMI+Inferior AMI). Of the 479 pts,78 were not suitable for thrombolysis,16 Killip class > 2,14             were sub‐acute AMi,30 pts were receiving ACE‐inhibitors at entry, and 51 were classified as unstable angina (no enzymatic variation). The study was adressed to             AMI thrombolysed and reperfused. The same was used in the 2° trial.

Dear Marco

I found the data on mortality with very dificult in old registry.

Mortality In reperfused  Captopril group (131 patients);  2 anterior AMI (1 Re‐IMA and 1 Heart Failure) during hospitalization, and 3 anterior AMI for HF  in the fisrt 6 month

                                  2 inferior AMI (Re‐IMA within 3 months)

                                   Place group: (128 patients); Anterior AMI; (2 Re‐IMA and 3 HF ) durng hospitalization; and 9 anterior AMI for HF   in the first 8 months.

                                    2 inferior AMI (re‐AMI within 6 months)

                                    Unreperfused AMI 31 pts,  captopril group: : 4 pts in the first 6 months for HF or Sudden Death

                                                                        placebo group; 4 pts in the first 6 months  for  irreversible HF.

The patients with unstable angina in both groups did not show mortality during the first year.  

Unfortunately I did not find  the data regarding the excluded patients from randomization.

Pietro Di Pasquale.

PS

I would like to inform that the 2° study was the first in the literature comparing the assocition of Beta‐blocker and ACE‐in with Beta‐blocker alone,captopril alone and placebo. The results of this study showed that i.v. beta‐blocker treatment in acute phases of AMI is not benefical compared with captopri,while as now recommended the benefical effects of beta blockers are reported when administerd later and orally.

From:

Eugenio Santoro [ mailto:eugenio@marionegri.it] 
 Sent: July 21, 2008 7:03 AM 
 To: maiperez@interchange.ubc.ca 
 Cc: Centro Studi ANMCO; franzosi@marionegri.it; maggioni@anmco.it 
 Subject: Fwd: RCT information 
   
 
 Dear Dr. Perez, 
   
 my name is Eugenio Santoro and I am co‐author of the paper: 
   
 “GISSI‐3. Effects of lisiriopril and transdermal glyceryl trinitrate singly and together on 6‐week mortality and ventricular function after acute myocardial infarction. The Lancet 1994 May 7;343(8906):1115‐22. 
 Pedrazzini G, Santoro E, Latini R, Fromm L, Franzosi MG, Mocetti T, et al. Causes of death in patients with acute myocardial infarction treated with angiotensin‐converting enzyme inhibitors: findings from the Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto (GISSI)‐3 trial. Am Heart J 2008 Feb;155(2):388‐94.” 
 
 I was also involved in the statistical analyses of the GISSI‐3 main papers. 
   
 This is to answer to your message related to the 10 July 2008 request of receiving data for the systematic review entitled  "Effect of early treatment with anti‐hypertensive drugs on short and long‐term health outcomes in patients with an acute cardiovascular event ". 
   
 Firstly the discrepancies you found about the 42 days mortality rate between the original (1994) and the 2008 paper was due to the fact that some extra mortality data was found. This has been already reported and well documented in the 1996 paper (GISSI3. Six month effects of early treatment with lisinopril and .. JACC 1996;27:337‐44)  when, performing the 6 months follow‐up analyses, we traced mortality data (at 6 weeks and 6 months) of 423 randomized patients originally excluded from the 1994 paper because of the lack of any kind of data. 
   
 This also explains why the number of patients analyzed for the 6 weeks mortality in the 1994 paper were 18895 while they were 19318 in the 1996 (and 2008) paper. 
   
 In conclusion, the correct data are 
   
 Mortality at week 6  from 2008 paper 
 Lisinopril group: 619 (6.4%) out of 9646 pts 
 Control group (no Lisinopril): 693 (7.2%) out of 9672 
 
   
 Coming back to your request these are the 10 days mortality data stratified by treatment allocation: 
   
   
 All cause mortality at 10 days 
 ‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐ 
 Lisinopril group: 435 (4.5%) out of 9646 pts 
 Control group (no Lisinopril): 486 (5.0%) out of 9672 
 
 Nitrate group: 443 (4.6%) out of 9663 pts 
 Control group (no Nitrate): 478 (4.9%) out of 9655 
 
 Please, dont’t esitate to contact me for any clarification. 
   
 Best wishes 
 
 Eugenio Santoro 
 on behalf of the GISSI group

eugenio

From: Eugenio Santoro [ mailto:eugenio@marionegri.it] 
 Sent: April 8, 2009 7:38 AM 
 To: MARCO PEREZ 
 Subject: GISSI‐3 Mortality day 2: RCT information 
   
 
 Dear Dr. Perez, 
 
 I apologize for the very long delay. 
 
 Please find enclosed the requested data. 
 
 Consider that I have used deaths at day 0 (pts dead the same date they entered into the trial) and death at day 1 (the day after randomization) as mortality at day 2. 
 
 The results, stratified by treatment allocation are: 
 
 All cause mortality at  day 2 
 ‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐ 
 Lisinopril group: 200 (2.1%) out of 9646 pts 
 Control group (no Lisinopril): 234 (2.4%) out of 9672 
 
 Nitrate group: 196 (2.0%) out of 9663 pts 
 Control group (no Nitrate): 238 (2.5%) out of 9655 
   
 Sorry again for the dealy 
 
 Please don't hesitate to contact me for any clarification. 
 
 I will be very grateful if you could notify me when the paper is published. 
 
 Best wishes 
 
 Eugenio Santoro

Appendix 3. Glossary & Definitions

ACC‐American College of cardiology

ACE‐I ‐ Angiotensin converting enzyme inhibitors

AHA‐American heart association

AMI‐ Acute myocardial infarction

ARBs ‐Angiotensin II receptor blockers

ARR‐Absolute risk reduction

ASA‐American Stroke association

AT1‐angiotensin II type 1 receptor

BB‐ Beta‐adrenergic receptor blockers

BHS‐British hypertension society

BPM‐Beats per minute

CCB‐ Calcium channel blockers

CHF‐Congestive heart failure

CK‐Creatine Kinase

CPK‐Phosphocreatine Kinase

CVE‐ Cardiovascular event

DBP‐Diastolic blood pressure

EOF: End of follow‐up

ESC‐European society of cardiology

ESH‐European society of hypertension

FDA‐Federal drug administration

HR‐Heart rate

ICD‐ International classification of diseases

ICH‐International conference of Harmonization

ISH‐International society of hypertension

IU‐International units

IV‐Intravenous

JNC‐Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure.

LVFP‐Left ventricular filling pressure

MAP‐ Mean arterial pressure

MESH‐Medical subject headings

Mo‐Month

N/A‐ Not applicable

NA‐ Not available

NNH‐Number needed to harm

NNT‐Number needed to treat

NO‐Nitric oxide

NR‐ Not reported

NS‐Not significant

NSTEMI‐Non‐ST elevation myocardial infarction

NTG‐Nitroglycerin

RAAS‐Renin angiotensin aldosterone system

RCT‐Randomized controlled trial

RR‐Relative risk

RRR‐Relative risk reduction

SAE‐Serious adverse events

SBP‐Systolic blood pressure

SD‐Standard deviation

SE‐ Standard error

SNP‐Sodium nitroprusside

UA‐Unstable angina

WHO‐World Health Organization

Data and analyses

Comparison 1. Nitrates.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 All‐cause mortality at 2 days	6	82624	Risk Ratio (M‐H, Fixed, 95% CI)	0.81 [0.74, 0.89]
2 All‐cause mortality at 10 days	16	84185	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.86, 0.96]
2.1 immediate treatment	10	6007	Risk Ratio (M‐H, Fixed, 95% CI)	0.84 [0.69, 1.01]
2.2 short‐term treatment	6	78178	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.86, 0.97]
3 All‐cause mortality at ≥30 days	10	6341	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.81, 1.02]
3.1 immediate treatment	7	5771	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.82, 1.04]
3.2 short‐term treatment	3	570	Risk Ratio (M‐H, Fixed, 95% CI)	0.72 [0.48, 1.10]
4 Weighted mean change in Systolic Blood pressure during first 24 hours	7	1958	Mean Difference (IV, Fixed, 95% CI)	‐12.67 [‐14.51, ‐10.83]
5 Weighted mean change in Diastolic Blood pressure during first 24 hours	6	1146	Mean Difference (IV, Fixed, 95% CI)	‐7.50 [‐9.07, ‐5.93]
6 Weighted mean change in heart rate during first 24 hours	6	810	Mean Difference (IV, Fixed, 95% CI)	‐0.83 [‐2.83, 1.17]
7 Sensitivity analysis mortality 0‐2 vs. 3‐10 days vs. day 11 to end of treatment (35‐42 days)	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
7.1 mortality up to day 2	2	77368	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.74, 0.90]
7.2 mortality from day 3 to day 10	2	75792	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.91, 1.06]
7.3 mortality from day 11 to 35‐42 days	2	73319	Risk Ratio (M‐H, Fixed, 95% CI)	1.10 [1.00, 1.22]

1.1. Analysis.

Comparison 1 Nitrates, Outcome 1 All‐cause mortality at 2 days.

1.2. Analysis.

Comparison 1 Nitrates, Outcome 2 All‐cause mortality at 10 days.

1.3. Analysis.

Comparison 1 Nitrates, Outcome 3 All‐cause mortality at ≥30 days.

1.4. Analysis.

Comparison 1 Nitrates, Outcome 4 Weighted mean change in Systolic Blood pressure during first 24 hours.

1.5. Analysis.

Comparison 1 Nitrates, Outcome 5 Weighted mean change in Diastolic Blood pressure during first 24 hours.

1.6. Analysis.

Comparison 1 Nitrates, Outcome 6 Weighted mean change in heart rate during first 24 hours.

1.7. Analysis.

Comparison 1 Nitrates, Outcome 7 Sensitivity analysis mortality 0‐2 vs. 3‐10 days vs. day 11 to end of treatment (35‐42 days).

Comparison 2. ACEi.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 All‐cause mortality at 2 days	3	77414	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.82, 1.00]
2 All‐cause mortality at 10 days	12	84456	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.87, 0.98]
2.1 immediate treatment	2	145	Risk Ratio (M‐H, Fixed, 95% CI)	0.68 [0.12, 3.98]
2.2 short‐term treatment	10	84311	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.87, 0.98]
3 Sensitivity analysis: mortality 0‐2 vs. 3‐10 days vs day 11 to end of treatment (35‐42 days)	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
3.1 mortality up to day 2	2	77368	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.82, 1.00]
3.2 mortality from 3 to 10 days	2	75792	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.86, 1.01]
3.3 mortality from 11 to 35‐42 days	2	73345	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.85, 1.03]

2.1. Analysis.

Comparison 2 ACEi, Outcome 1 All‐cause mortality at 2 days.

2.2. Analysis.

Comparison 2 ACEi, Outcome 2 All‐cause mortality at 10 days.

2.3. Analysis.

Comparison 2 ACEi, Outcome 3 Sensitivity analysis: mortality 0‐2 vs. 3‐10 days vs day 11 to end of treatment (35‐42 days).

Comparison 3. BB.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 All‐cause mortality at 2 days	6	68007	Risk Ratio (M‐H, Fixed, 95% CI)	0.95 [0.85, 1.07]
2 All‐cause mortality at 10 days	20	72600	Risk Ratio (M‐H, Fixed, 95% CI)	0.96 [0.91, 1.02]
2.1 immediate treatment	6	1143	Risk Ratio (M‐H, Fixed, 95% CI)	1.12 [0.60, 2.07]
2.2 short‐term treatment	14	71457	Risk Ratio (M‐H, Fixed, 95% CI)	0.96 [0.91, 1.02]
3 All‐cause mortality at ≥30 day	6	18481	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.84, 0.99]
3.1 immediate treatment	1	108	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.2 short‐term treatment	5	18373	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.84, 0.99]
4 Weighted mean change in Systolic Blood pressure during first 24 hours	6	738	Mean Difference (IV, Fixed, 95% CI)	‐12.54 [‐15.63, ‐9.45]
5 Weighted mean change in Diastolic Blood pressure during first 24 hours	6	738	Mean Difference (IV, Fixed, 95% CI)	‐3.35 [‐5.43, ‐1.28]
6 Weighted mean change in heart rate during first 24 hours	5	594	Mean Difference (IV, Fixed, 95% CI)	‐9.68 [‐11.99, ‐7.37]
7 Sensitivity analysis: mortality at 2 days, according to trial design	6	68007	Risk Ratio (M‐H, Fixed, 95% CI)	0.95 [0.85, 1.07]
7.1 Double blind trials	4	51814	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.91, 1.19]
7.2 Not double‐blind trials	2	16193	Risk Ratio (M‐H, Fixed, 95% CI)	0.73 [0.58, 0.91]

3.1. Analysis.

Comparison 3 BB, Outcome 1 All‐cause mortality at 2 days.

3.2. Analysis.

Comparison 3 BB, Outcome 2 All‐cause mortality at 10 days.

3.3. Analysis.

Comparison 3 BB, Outcome 3 All‐cause mortality at ≥30 day.

3.4. Analysis.

Comparison 3 BB, Outcome 4 Weighted mean change in Systolic Blood pressure during first 24 hours.

3.5. Analysis.

Comparison 3 BB, Outcome 5 Weighted mean change in Diastolic Blood pressure during first 24 hours.

3.6. Analysis.

Comparison 3 BB, Outcome 6 Weighted mean change in heart rate during first 24 hours.

3.7. Analysis.

Comparison 3 BB, Outcome 7 Sensitivity analysis: mortality at 2 days, according to trial design.

Comparison 4. CCB.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 All‐cause mortality at 2 days	3	242	Risk Ratio (M‐H, Fixed, 95% CI)	2.33 [0.62, 8.78]
2 All‐cause mortality at 10 days	16	1988	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.74, 1.40]
2.1 immediate treatment	1	88	Risk Ratio (M‐H, Fixed, 95% CI)	3.0 [0.13, 71.70]
2.2 short‐term treatment	15	1900	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.73, 1.38]
3 All‐cause mortality at ≥30 days	2	198	Risk Ratio (M‐H, Fixed, 95% CI)	1.47 [0.58, 3.68]
3.1 immediate treatment	1	108	Risk Ratio (M‐H, Fixed, 95% CI)	1.4 [0.47, 4.14]
3.2 short‐term treatment	1	90	Risk Ratio (M‐H, Fixed, 95% CI)	1.64 [0.29, 9.35]
4 Weighted mean change in Systolic Blood pressure during first 24 hours	7	755	Mean Difference (IV, Fixed, 95% CI)	‐5.49 [‐8.42, ‐2.56]
5 Weighted mean change in Diastolic Blood pressure during first 24 hours	6	665	Mean Difference (IV, Fixed, 95% CI)	‐5.08 [‐7.00, ‐3.15]
6 Weighted mean change in heart rate during first 24 hours	5	420	Mean Difference (IV, Fixed, 95% CI)	‐1.10 [‐4.60, 2.40]
7 Sensititity analysis: mortality at 10 days, according underlying CVE	15	1946	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.73, 1.41]
7.1 acute myocardial infarction	10	935	Risk Ratio (M‐H, Random, 95% CI)	1.60 [0.90, 2.86]
7.2 acute stroke	5	1011	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.54, 1.21]
8 Sensitivity Analysis: mortality at 10 days, according sub‐class of CCB in AMI	10	935	Risk Ratio (M‐H, Random, 95% CI)	1.60 [0.90, 2.86]
8.1 Dihydropyridines	4	577	Risk Ratio (M‐H, Random, 95% CI)	1.91 [0.98, 3.72]
8.2 Non‐dihydropyridines	6	358	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.28, 3.00]

4.1. Analysis.

Comparison 4 CCB, Outcome 1 All‐cause mortality at 2 days.

4.2. Analysis.

Comparison 4 CCB, Outcome 2 All‐cause mortality at 10 days.

4.3. Analysis.

Comparison 4 CCB, Outcome 3 All‐cause mortality at ≥30 days.

4.4. Analysis.

Comparison 4 CCB, Outcome 4 Weighted mean change in Systolic Blood pressure during first 24 hours.

4.5. Analysis.

Comparison 4 CCB, Outcome 5 Weighted mean change in Diastolic Blood pressure during first 24 hours.

4.6. Analysis.

Comparison 4 CCB, Outcome 6 Weighted mean change in heart rate during first 24 hours.

4.7. Analysis.

Comparison 4 CCB, Outcome 7 Sensititity analysis: mortality at 10 days, according underlying CVE.

4.8. Analysis.

Comparison 4 CCB, Outcome 8 Sensitivity Analysis: mortality at 10 days, according sub‐class of CCB in AMI.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Beaufils 1988.

Methods	Multi‐centre ( 30‐French) Double‐blind Method of randomization: Not reported (NR) Concealment of allocation: NR Duration of treatment: 10 days Follow‐up: 10 days
Participants	303 patients within 6 hours after onset of symptoms of first AMI Inclusion criteria: Involved two stages: Patients age less than 75 years, suffering from a first myocardial infarct* were pre‐included and randomized. Confirmation of inclusion was with increase of CK‐MB enzymes levels. * chest pain lasting more than 15 minutes , persistent abnormalities of repolarization. Exclusion criteria: Patients with recurrent infarctions, in shock, un‐interpretable ECG, or those who had received electrocardioversion, some form of nitrates, beta‐blockers, CCB, or thrombolytic were excluded. Attrition data: Screened: NR Total randomized patients: 303 (270), Molsidomine: n= 152 (133), Placebo: n= 151 (137) Figures in brackets reflect the patients included in the mortality analysis. Not an intention to treat analysis. Total withdrawals (discontinuation of drug): 24, Molsidomine  10, Placebo 14 Withdrawals due to adverse events: 22, Molsidomine  9, Placebo 13 Total lost to follow‐up: Not included in the analysis 33/ 303 (10 %) [ non confirmation of AMI diagnosis 19, recurrent infarction 3, after 6 hour 7, used of nitroglycerin 4]. Molsidomine: 19 Placebo: 14 Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years), molsidomine (M): 55.4± 9.9, placebo (P): 57.1±10.3 Sex (M/ F), M: 114/19, P: 124/13 Treatment time: M: 3.57 ±2, P: 3.45 ±1 Infarct location: anterior / inferior, M: ( 57 / 73), P:  (55/ 80 ) BP (mm Hg), M: 133/ 85, P: 129/83 Medical history: NR
Interventions	Molsidomine vs. Placebo: Drug regimen: Molsidomine: Initially 2 mg of an oral form and 2 mg of a sublingual form.  Total 16 mg on the first day, 12 mg on the second day, (2 mg x 6 times a day), and 6 mg (2 mg tid) from third day to day 10 Placebo: identical form (no further details) Other interventions: NR
Outcomes	Mortality:  obtained from text, page 130 Molsidomine: 2day:NR ; 10 day: 6/133 (eof)  Placebo: 2day:NR ; 10 day: 11/137  (eof) Total non‐fatal SAE: NR Blood Pressure data: NR Heart rate data: NR
Notes	Funding: Hoechst Laboratories, Paris Dates of conducting the trial : NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"allocated at random" no further details"
Allocation concealment?	Unclear risk	Not reported. probably done as it was a double‐blind trial
Blinding? All outcomes	Low risk	double‐blind
Incomplete outcome data addressed? 2‐day mortality	High risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Unclear risk	33/303 (10%) Not included in the analysis; Molsidomine: 19; Placebo: 14
Incomplete outcome data addressed? 30‐day mortality	High risk	Not reported
Free of selective reporting?	Low risk	Reporting according to methods

Branagan 1986.

Methods	Single‐site study (Ireland) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 3 days Follow‐up: 1 month
Participants	121 patients with suspected AMI within 6 hr of onset Inclusion criteria: Patients within 6 hours of suspected AMI (acute chest pain lasting more than fifteen minutes) Exclusion criteria: Over 70 years of age, women capable of child‐bearing, heart rate greater than 110 per min, or less than 60 per min., complete heart block, heart failure requiring diuretic therapy or mechanical support. SBP less than 85 mm Hg, ventricular fibrillation, or were on nifedipine. Attrition data:  Screened: N = 1220, total randomized patients, Nifedipine (N ): =60, Placebo (P): n= 61 Total lost of follow‐up (randomized but not considered for analysis in the original publication): Nifedipine:  6 (10%, because of non‐ischaemic chest pain), placebo: 7 (11%, because of erroneously entered 1, non‐ischaemic pain 6) Thus, the total of patients analyzed (mortality) in the original publications was Nifedipine (N ): n=54, Placebo (P): n= 54 Total withdrawals (discontinuation of drug): N: 8 (15%), P: 7 (13%) Withdrawals due to adverse events: Nifedipine: 7 (13%, bradicardia 2, ventricular tachycardia 1, complete heart block 1, fatal cardiogenic shock 1, fatal heart failure 2) Placebo: 7 ( 13%, bradicardia 3, CABG 2, fatal cardiogenic shock 2) Baseline characteristics: Note: continuous variables are expressed as mean ± SE Male/female N:  37/9, P: 44/8 Age (years) : N:55.6 ± 1.3, P: 56.6±1.25 Time to treatment (hours): N: 3.33±0.2, P: 3.28±0.18
Interventions	Nifedipine vs. Placebo Drug regimen: Nifedipine 10 mg sublingually , after 4 hours  a second capsule was administrated, and then  one capsule every 6 hours for 48 hours. Placebo, no further details. Other interventions: NR
Outcomes	Mortality:  obtained it from text, page 863 N: at 2 day: NR; 10 day: NR; eof (1m): 7/54 P: at 2 day: NR; 10 day: NR; eof (1m):5/54 Total Non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR Heart rate change during first 24 hours: NR
Notes	Funding: Bayer  UK. Dates of conducting the trial : July 1982‐ September 1984
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomly allocated" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 30‐day mortality	Low risk	There was a 10% lost to follow‐up
Free of selective reporting?	Low risk	Reporting according to methods

Bussmann 1981.

Methods	Single‐centre ( West Germany). Open label Method of randomization/: NR Concealment of allocation: NR Duration of treatment:  48 hours Follow‐up: 18 months
Participants	60 patients with AMI within 24 hours Inclusion criteria: Patients with acute myocardial infarction (prolonged chest pain, ST‐segment elevation with development of q waves, and increase in CK and CK‐MB activity), diastolic pulmonary arterial or pulmonary capillary pressure was over 15 mm Hg. Two subgroups: early < 8 hours; and late > 8 hours Exclusion criteria: Not stated Attrition data: Screened: NR, Total randomized patients: Total N= 60, Nitroglycerin (N): n=31, Control (C): n=29 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: 3 patients were Lost of follow‐up during 1.5 years Baseline characteristics: Mean Age (years): N: 59±11, C: 63±11 Anterior/posterior infarct : N: 17/14, C:15/14 Time to intervention (hours):  N: 10.0, C: 10.4 HR (bpm): N: 88±19, C:87±16 MAP (mm Hg): N: 108±19, C:109±15
Interventions	Nitroglycerin vs. No treatment  ( Control) Drug regimens: Nitroglycerin: Two subgroups: Early: (NTG  given within 8 h)= 0.75‐ 6 mg/ h (12.5‐100 mcg/min) iv infusion according to DPAP  and BP (mean dose reported was from 2.8  to3.2 mg/h)  (50 mcg/min) for 48 h Late: (NTG given after 8 h) same dose regimen (mean dose reported was from 2.7  to3.6 mg/h) for 48 h Control (29)= not stated method Other interventions: Nitroglycerin group:  Lidocaine 4, digitalis 8, furosemide 3, atropine 5, antihypertensive 2, morphine 12 Control group: Lidocaine 8, digitalis 11, furosemide 11, atropine 11, antihypertensive 5, morphine 19
Outcomes	Mortality  obtained from text, page 418 & 419.[Klinische‐Wochenschrift.1983;61(8):417‐22] Nitroglycerin: 2d:NR ; 10d:2/31 (6.45% ) ; eof(3 m.):6/31 (19%)  Control:2d:NR ; 10d:5/29 ( 17%)  ; eof(3 m.):12/29 (41%) Non‐fatal SAE : NR Blood pressure data: NR There was only a report for mean arterial pressure in graph on page 620. Heart rate: obtained from graph & text on page 620 (Circulation 1981; vol 63) The calculated weighted mean HR change at 24 h was: Nitroglycerin: 1.76 ±16 Control:‐2.2 ±16
Notes	Funding: NR Dates of conducting the trial : NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were assigned randomly" no further details
Allocation concealment?	Unclear risk	Not reported,probably not done
Blinding? All outcomes	High risk	Not reported, most likely not done
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Low risk	3 patients were Lost to follow‐up during 1.5 years
Free of selective reporting?	Unclear risk	exclusion criteria not reported,

Bussmann 1992.

Methods	Single‐site study (West Germany) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 2 days Follow‐up: 2 days
Participants	46 patients within 18 hours after onset of symptoms of AMI Inclusion criteria: Patients with acute myocardial infarction within 2‐18 h Exclusion criteria:Patients were excluded if Infarct was over 18 h, or decreasing CK  concentrations, Systolic < 100 mmHg, cardiogenic shock Attrition data: Screened: NR (40 % of the screened people were not able to enter the trial due to technical or organizational reason) Total randomized patients: 46 ; Captopril: n= 22, Placebo : n= 24  Total withdrawals (discontinuation of drug): NR  Withdrawals due to adverse events: Captopril: 2, Placebo : 1  Total lost to follow‐up: NR  Baseline characteristics: Age (years) : captopril (C): 62, Placebo (P): 59.9 Male/female: C: 19/3, P:18/6 Anterior/posterior infarct: C: 11/11, P:12/12 HR (bpm): C: 83.1, P:77.6 MAP (mm Hg) C: 100.1, P:93.9
Interventions	Captopril vs. Placebo  Drug regimen: Captopril: slow 2.5‐5 mg IV bolus injections followed by a continuous infusion of 1.5‐2.0 mg/ hour for 48 hours Placebo: No further details  Other interventions: Captopril:  Thrombolysis 11, nitrates 21, morphine 19 Placebo: Thrombolysis 12, nitrates 23, morphine 15
Outcomes	Mortality:  (obtained from text, page 653) Captopril:(n= 22); at day 2: 0, at day 10: 1, at EOF: NR  Placebo : (n= 24), at day 2: 0 , at day 10: 2, at EOF: NR  Total non‐fatal SAE: NR Blood Pressure; Reported as change in MAP, page 654 Heart rate: NR
Notes	Funding: NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomized" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	claimed to be a double‐blind trial
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods.

Charvat 1990.

Methods	Single‐centre (Kuwait), Open‐label (placebo‐controlled trial) Method of randomization: NR Method of allocation: NR Duration of treatment: 48 hours Follow‐up: 72 hours
Participants	129 patients within 6 hours of onset of suspected AMI* * Chest pain with 20 minutes or longer duration, ST elevation greater than 1 mm in 2 leads (40 ms after J point), cardiac enzymes. Attrition data: Screened: NR, Total randomized patients: Nitroglycerin (N): n=62, Placebo (P): n= 67 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: Mean age: (years): Nitroglycerin (N): N:52±8.25, Placebo (P): P:50±9.72 Mean baseline of BP: (mm Hg): N:140/82 ± 29.27/26.16, P:136/91 ± 32.44/23.66 Localization of AMI N: anterior 42, inferior 20 P: anterior 34, inferior 33 Previous Medical history: Nitroglycerin: diabetes 21, smoking 41, hypercholesterolemia 13, prior angina 30, prior MI 13. Placebo: diabetes 14, smoking 43, hypercholesterolemia 16, prior angina 18, prior MI 10.
Interventions	Nitroglycerin vs. Placebo Dose regimen Nitroglycerin: two subsets: a)     if BP <160/105 : Intravenous infusion of 10 mcg/min (to be increase 10 mcg/min  every 5 minutes) until SBP reached  20 mm Hg reduction but not lower than 100 mm Hg. b)     If > 160/105 : Intravenous infusion of 40 mcg/min (to be increase 10 mcg/min  every 5 minutes) until SBP reached  20 mm Hg and DBP <  90  mm Hg. Placebo; infusion of  5 % dextrose. (no further details) Treatment started within 6 hours of onset of pain and lasted for  48 hours   after randomization. The mean dose of nitroglycerin to obtained target was 75 ± 44 mcg/min. Other interventions: Sublingual nitroglycerin, morphine 10‐20 IV, nasal oxygen, iv morphine if pain not relieved, continuous lidocaine 1‐4 mg/min. CCB and BB if new ECG changes, diuretics if CHF, iv heparin to all patients. No thrombolysis was mentioned.
Outcomes	Mortality: obtained from table II, page 51 N: 2d:NR ; 10d:5/62  ; eof:N/A (FU=72 hours) Placebo: 2d: NR ; 10d:5/67  ; eof: N/A (FU=72 hours) Total Non‐fatal SAE: NR Blood Pressure; Data was obtained from text, page 51(single point in time after titration (2‐6 h). The calculated weighted mean BP change was: Nitroglycerin: SBP‐21±21.2; DBP‐11±12.6 Placebo: SBP‐1±25 ; DBP‐4 ±19 Standard deviation of change was not reported but imputed from end point Heart rate The calculated weighted mean HR change was: Nitroglycerin: 2±19 Placebo: ‐3 ±18.24
Notes	Funding: NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were divided in random into 2 groups"; no further details
Allocation concealment?	Unclear risk	not reported,
Blinding? All outcomes	High risk	not done
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods.

Chiche 1979.

Methods	Single‐centre (France), open‐label Method of randomization/ allocation: NR Duration of treatment : 7 days Follow‐up: 28 days
Participants	95 patients with signs & symptoms of myocardial infarction, within 12 h of the onset.
Interventions	Nitroglycerin (N): n= 50 Placebo (P; dextrose 5 %): n= 45 Dose regimen: Nitroglycerin: Intravenous infusion at initial rate of 10‐15 mg/min with progressive adjustments of 10‐15 mg/min according to clinical observation ( no more than 20 mm Hg drop of blood pressure); for 7 days Placebo: glucose 5% ( no futher details) Duration of treatment: seven days Mean dose of nitroglycerin was :50 mcg/min
Outcomes	Mortality: obtained from text, page 686 & 699 Nitroglycerin: 2d:NR ; 10d:NR ; eof(1m): 3/50 Placebo: 2d:NR ; 10d:NR ; eof(3 m): 8/45 Non‐fatal SAE: (page 693 ) LTA: Nitroglycerin: 2d:NR ; 10d:1/50 ; eof(1m):NR Placebo: 2d:NR ; 10d:3/45;eof(3 m):NR Blood pressure: No data was reported HR: NR
Notes	Funding: NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"Patients were randomized", no further details
Allocation concealment?	Unclear risk	Not reported
Blinding? All outcomes	High risk	Not done
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Free of selective reporting?	Low risk	Mortality data reported according to methods

Clausen 1966.

Methods	Single‐site study (Denmark) Open‐label Method of randomization: NR Concealment of allocation: selection was performed by one to the secretaries and no physician concerned took any part in the selection. Duration of treatment: 14 days Follow‐up: 22 days
Participants	130 patients within 24 hours after onset of symptoms of AMI Inclusion criteria: Diagnosis of AMI was based on ECG and enzymes.   Exclusion criteria: History of asthma   Attrition data: screened: Total randomized patients: 130, propranolol: n= 66   placebo: n= 64, Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: nil Baseline characteristics: Not reported based on the total of randomized patients. Medical history (%) Propranolol (n=53): diabetes 6(11), previous infarction 13(25), angina pectoris 35(66) No treatment: (n=57): diabetes 3(5), previous infarction 7(12), angina pectoris 34(60)
Interventions	Propranolol vs No treatment   Drug regimen: Propranolol: 10 mg orally four times in twenty‐four hours from day 1 to day 14. sometimes patients required propranolol IV 5 mg instead of the initial 10 mg.   Other interventions: All patients were given digitalis, diuretics metaraminol or procainamide if necessary.
Outcomes	Mortality:  obtained from table V, page 922 . Propranolol group: 2day:NR; 10day: 13/66  ; eof(21day,N/A ): 18/66 No treatment group: 2day:NR ; 10day: 13/64  ; eof(21 day,N/A ): 16/64 Total non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR   Heart rate change during first 24 hours: NR
Notes	Funding: Danish League Against Heart Disease Dates of conducting the trial : November 1965‐1966
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomly treated" no further details
Allocation concealment?	Unclear risk	selection was performed by one to the secretaries
Blinding? All outcomes	High risk	open control
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	High risk	Baseline characteristics: Not reported based on the total of randomized patients.

Cohn 1982.

Methods	Multi‐centre (11 in US). Double blind, Concealment of allocation: opaque code envelope Method of randomization:  Patients stratified within each hospital according to age (below or above 65), systolic BP (greater or lower than 150 mm Hg) and left ventricular filling pressure (greater or lower than 20 mmHg). Duration of treatment: 48 hours Length of follow‐up: 13 weeks
Participants	812 men within 24 hours of the onset of acute myocardial infarction* and with LVFP > 12 mm Hg Inclusion criteria: Patients with acute myocardial infarction in the past 24 hours (* chest pain, ST‐segment elevation, q waves lasting at least 0.04 sec or a new conduction defect) plus left ventricular filling pressure > 12 and SBP > 100 mm Hg). Exclusion criteria: Shock, hypertension likely to require specific anti‐hypertensive therapy, severe bronchopulmonary disease or other medical illness likely to prevent survival for 13 weeks Attrition data: Screened: 3,663, excluded on screening: 1577 the inclusion criteria were not met, 436 refused consent, and 838 rejected the use of catheter. Total randomized patients: Nitroprusside (N): n= 407, Placebo (P): n= 405. Total withdrawals (discontinuation of drug): N: 1.2 %, P:2% Withdrawals due to adverse events: NR Total lost to follow‐up: nil Baseline characteristics: Age ‐ years: Nitroprusside (N): 58.3, Placebo (P): 58.7 SBP(mmHg): N: 133.1, P:132.3 Neither DBP nor standard deviation of SBP at baseline is reported Time of infarction (h): N: 16.5, P:15.4 Anterior involvement (%): N: 50.4, P:48.1 Medical history (%): Nitroprusside: previous AMI 39.7, angina 53.4, hypertension 41.0, CHF 19.2, cerebrovascular accident 9.6 Placebo: previous AMI 45.2, angina 61, hypertension 40.7, CHF 17.8, cerebrovascular accident 8.7 Medication in previous 2 months (%) N:  Diuretics 29.5, nitrates 33.3, digitalis 19.2, beta blockers 11.8 P: Diuretics 28.5, nitrates 40.2*, digitalis 19.3, beta blockers 16.3
Interventions	Nitroprusside vs. Placebo Drug regimen: Nitroprusside:  IV infusion was started at a rate of six drops per  minute ( 10 mcg) and increase 10 mcg/min q 5 min until maximum dose allowed or  achievement of the target* or a side effect happen. The maintenance treatment was for 48 h Placebo:  (dextrose solution) same regimen as nitroprusside *Target: to attain SBP reduction of 20% of BL +76 mm Hg mean dose reported was 72.8 mcg/h at titration mean dose reported was 92.3 mcg/h at 24 h mean dose reported was 94.4 mcg/h) at 48 h Other interventions: Except for diuretics No antihypertensive was allowed. Nitrates were not allowed either. Beta blockers were not given  except in unusual circumstances and previous beta‐blockers was usually discontinued before randomization. Analgesics, antiarrhythmic drugs, diuretics, digitalis and anticoagulants were used according to local policies. During first 24 hours the use of diuretics was considerably higher in the placebo group. In the first 8 hours 45 % of patients given placebo received diuretics as compared with 11 % of patients given nitroprusside.
Outcomes	Mortality:  (table 2‐page 1132; graph fig. 2 page 1134; table 3‐ p.1133 ) Nitroprusside: 2d:11/407 ; 10d: 29/407 ; eof(3m): 69/407  Placebo:          2d:9/405;    10d: 22/405  ; eof(3m):77/405 Non‐fatal SAE: NR Blood pressure: Data was obtained from graphs (fig 1 on page 1132). The calculated weighted mean BP change was: Nitroprusside: SBP ‐16.13± NR ; DBP: NR Placebo: SBP‐6.55± NR ; DBP: NR Heart rate Data was obtained from graphs (fig 1 on page 1132). The calculated weighted mean HR change was: Nitroprusside : 3.93 ± NR Placebo: 1.86 ± NR
Notes	Funding: Cooperative Studies Program of the Medical Research Service, Veterans Administration Central Office, Washinton, DC. Dates of conducting the trial : 1975
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Patients were randomized using stratified randomization within each hospital according to age (below or above 65), systolic BP (greater or lower than 150 mm Hg) and left ventricular filling pressure (greater or lower than 20 mmHg). Block of four randomization within each stratum
Allocation concealment?	Low risk	use of  an opaque code envelope
Blinding? All outcomes	Low risk	Drug solutions were diluted with tinted syringes and infused intravenously from foil‐wrapped plastic bags in a double‐blind fashion. The code was not available at the hospital
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Free of selective reporting?	Low risk	Data reported according to methods

COMMIT 2005.

Methods	Multi‐centre ( 1, 250‐China) Double‐blind Method of randomization: NR Concealment of allocation: Sequentially‐numbered sealed treatment packs prepared centrally Duration of treatment: 28 days Follow‐up: 28 days
Participants	45,852 patients within 24 hours after onset of  AMI   Inclusion criteria: Patients who presented with ST elevation, left‐bundle branch block or ST depression within 24 hours.   Exclusion criteria: Not to have a clear indication for or contraindication to any of the study treatments. Patients scheduled for primary percutaneous coronary intervention, other life‐threatening disease, high risk of adverse effects with metoprolol ( SBP < 100 mm Hg) or low heart rate (<50 bpm) heart block, or cardiogenic shock. Note: evidence of moderate heart failure (killip II or III) was not an exclusion criterion.   Attrition data: Screened: NR, Total randomized patients:45,852, Metoprolol: =22,929, Placebo : n= 22,923 Total withdrawals (discontinuation of drug):  Metoprolol: 3024 (13.2%), Placebo: 1836 (8.0%) Total withdrawals due to adverse events: NR Total lost to follow‐up: 2 (1 in metoprolol group, 1 placebo group) Baseline characteristics: Confirmed MI: 95.8% (Metoprolol [M] 21,993; Placebo [P] 21,955) Possible MI: 1.8% (M 405, P 409) Unstable angina 1.3% (M 294, P 302)  Other 1.1% (M 236, P 256)  Note: continuous variables are expressed as mean ± SD Age (years): Metoprolol (M):61.4±11.8, Placebo (P): 61.3±11.8 Female : M: 6431(28 %), P:  6328(27.6 %) Time since onset (h): M: 10.3 ±6.7, P: 10.3 ±6.7 SBP (mm Hg): M: 128.2 ±22.6, P:  128.2 ± 22.5 Heart rate: M: 82 ±17.3, P:  82.3 ±17.1 ECG findings (%) Metoprolol:  ST elevation (86.7), bundle branch block (6.2), ST depression without ST elevation (7.1) Placebo: ST elevation (86.8), bundle branch block (6.5), ST depression without ST elevation (6.7) Killip class I; II; III (%): M: 75.3; 19.9; 4.7 , P:  75.6;19.8; 4.6 Medical history(%) Metoprolol: Previous MI (8.4), hypertension (43.4), Placebo: Previous MI (8.3), hypertension (43.1),
Interventions	Metoprolol vs. Placebo Note: there was another comparison, aspirin plus clopidogrel vs. aspirin, as per a 2x2 factorial design. This comparison is not discuss in this review further Drug regimen: Metoprolol: Initially, 5 mg IV over 2‐3 min, x 3 doses (every 2‐3 minutes if the SBP>90  mm Hg, and > 50 bpm). Then, 15 minutes after these IV doses,  a 50 mg orally was given every 6 hours during days 0‐1. From day 2 onwards a 200 mg controlled‐release for up to 4 weeks Placebo: Same regiment, No further details The three injections were received by 90% of those allocated to metoprolol and 96% allocated to placebo. The oral treatment was completed in 86.2% for those in metoprolol group and 91.6% in the placebo group. Other interventions: Non‐study b‐blockers and anti‐platelets were to be avoided. Unless contraindicated thrombolysis was administrated to all patients before randomization Metoprolol group: non‐trial beta blocker (6.1%)*, fibrinolytic (54.3%), anticoagulant (74.4%) antiarrhythmic (22.0%), ACE inhibitor (67.2 %)*, nitrate (94.1%), diuretic (24.2 %), CCB (10.9%). Placebo group: non‐trial beta blocker (15.7%)*, fibrinolytic (54.6%), anticoagulant (74.7%) antiarrhythmic (22.7%), ACE inhibitor (69.3%)*, nitrate (94.3%), diuretic (22.4%), CCB (12.6%). * p < 0.0001
Outcomes	Mortality: obtained from figure 3, page 1626  Metoprolol: n=22,929 2day: 395 (1.7%); 10(7)day:1441(6.28%);eof(4wk,N/A):1774(7.7%) Placebo : n= 22,923 2day: 364 (1.6%); 10(7)day:1449 (6.32%); eof (4wk, N/A): 1797(7.8%) Total non‐fatal SAE: NR Blood Pressure change during first 24 h: NR Note: Fatality rate according to baseline BP:                 < 120; 120‐139, 140‐159, > 160 Metoprolol: 9.6,       6.8,         6.5,        7.2 Placebo:     8.8,       7.6,         7.0,        7.0   The above data was obtained from table 6, page 1628.   Heart rate change during first 24 h: NR   Fatality rate according to baseline BP:                 < 70;    70‐89,      90‐109,    > 110 Metoprolol: 4.3,       6.0,         11.0,       20.3 Placebo:     4.1,       6.6,         10.7,       18.3
Notes	Funding: By Sanofi‐Aventis and Bristol‐Myers Squibb (clopidogrel manufacturers) and  by AstraZeneca (metoprolol manufacturers) Dates of conducting the trial: August 1999‐ February 2005 COMMIT trial is also the Second Chinese Cardiac Study (CCS‐2).
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"random allocation" centralized computer driven system
Allocation concealment?	Low risk	Sequentially‐numbered sealed treatment packs prepared centrally
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Data reported according to methods

CONSENSUS‐II 1992.

Methods	Multi‐centre (103 Scandinavian Countries) Double‐blind Method of randomization: Patients were stratified in blocks of 2‐10 according to whether they had had previous myocardial infarction and according to study centre. No further details Concealment of allocation: Package of the study drugs were labeled with numbers assigned to the patients Duration of treatment: 6 months Follow‐up: 6 months
Participants	6,090 patients within 24 hours after onset of symptoms of AMI   Inclusion criteria: Presented within 24 of the onset of chest pain suspected of AMI (elevation of the ST segment in two or more contiguous ECG leads; new pathologic Q‐waves or elevated plasma levels of enzymes) Exclusion criteria: Blood pressure < 105/ 65 mm Hg (supine), need for vasopressor, severe valvular stenosis, untreated third‐degree AV block, a history of angioedema, or sensitivity to ACE inhibitors, or the use of such drugs within one week before the infarction; severe renal, hepatic or hematologic disorders, history of cerebral transient ischemic attacks or other life‐threatening condition, a clear indication for ACE inhibitors   Attrition data: Screened: 10,387, Excluded by screening: 4297 (41%) Total randomized patients: 6090, Enalapril: n= 3044 , Placebo : n= 3046 Total withdrawals (discontinuation of drug): NR Total withdrawals due to adverse events: NR Total lost to follow‐up: NIL Baseline characteristics: Mean age (years): Enalapril (E): 65.7, Placebo (P): 65.8, No. of men (%): E: 2208 (73%), P: 2239 (74%) Time to randomization (hours): E: 15, P:  15 Inferior infarction E:  1006 (33 %), P: 1023(34 %) SBP/ DBP (mm Hg): E: 133 /80, P:  134/ 80 Heart rate (bpm): E: 75, P:75 Medical history (%) Enalapril: previous AMI 709 (23), heart failure 203 (7), diabetes 355 (12), angina pectoris 1141 (37), smoking 1135 (37), Placebo: previous AMI 723 (24), heart failure 171 (6), diabetes 330 (11), angina pectoris 1195 (39), smoking 1101 (36)
Interventions	Enalapril  vs. placebo Drug regimen: Enalaprilat: 1 mg IV diluted in 100 ml of 0.9 percent saline over 2 hours*. Then, 6 hours after  oral  enalapril at dose of 2.5 mg bid titrated up to 20 mg /day on the fifth day and thereafter for  month.  Placebo:  No further details * The infusion was discontinued if systolic/ diastolic blood pressure decrease to < 100 / 60 mm Hg. Other interventions: according to study groups Enalapril group: beta‐blockers 2020 (66%), CCB 715(23%), nitrates 1607(53%), thrombolytics 1702 (56 %) Placebo group: beta‐blockers 2053 (67%), CCB 716(24%), nitrates 1611(53%), thrombolytics 1712 (56 %)
Outcomes	Mortality:  obtained from text, page 679 Enalapril: 2day: NR; 10day: 140/3044(4.6%); end of trial: N/A; 312/ 3044 (10.2%); at 6 mo 334 (11%) Placebo: 2day: NR; 10day: 131/3046 (4.3%) ; end or trial: N/A; 286/ 3046 (9.4%); at 6 mo 310 (10.2%) Total non‐fatal SAE: NR Blood Pressure change during first 24 h: NR Heart rate change during first 24 h: NR
Notes	Funding: Merck Sharp and Dohme Dates of conducting the trial: March 1990‐ March 1991
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Patients were stratified in blocks of 2‐10 according to whether they had had previous myocardial infarction and according to study centre.
Allocation concealment?	Low risk	Package of the study drugs were labelled only with numbers assigned to the patients
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Data reported according to methods

Crea 1985.

Methods	Single‐site study (UK) Single‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: during hospital stay Follow‐up: until discharge of hospital
Participants	17 patients with AMI within 12 hours of onset Inclusion criteria: Patients with prolonged chest pain typical of acute myocardial infarction and with ECG changes consistent with transmural ischemia. Exclusion criteria: Older than 75 years, SBP < 90 mm Hg, heart rate less than 55 beats/min, severe hypertension requiring intravenous vasodilators, PR interval > 0.3 second, second and third‐degree AV block, QRS interval longer than 0,1 second, severe heart failure, syncope or cardiac arrest before admission. Attrition data: Total randomized patients: Verapamil (V): n=8 Placebo (P): n=9 Total withdrawals (discontinuation of drug): V= 2; P= 0 Withdrawals due to adverse events: V= 2; P= 0 Total lost of follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years) V: 59±7 P: 56±6 BP (mm Hg): V:118/81,P:127/77 Previous AMI V: 2(25%) P: 1(11%) Time to treatment (hours) V: 6±3 P: 7±3 Anterior involvement V: 4(50%) P: 5(55%)
Interventions	Verapamil (V): n=8 Placebo (P): n=9 Drug regimen: V: verapamil IV bolus of 10 mg at interval of 30 min up to a total dose of 40 mg , followed by oral administration of 80 mg 3 times daily until discharge*. P: Saline solution and placebo tablets were administered in a manner similar to that of verapamil. * medication was interrupted if decrease in systolic blood pressure to less than 90 mm Hg and second or third‐degree AV block mean dose was not reported Other interventions: Nitroglycerin, and morphine were given for chest pain.
Outcomes	Mortality: obtained it from text, page 902 Verapamil: 2day: NR; 10day: 0/8; eof: N/A Placebo: 2day: NR; 10 day: 2/9; eof: N/A Total non‐fatal SAE: NR Individual SAE: CHF: V=2/8; P=2 /9 Second‐degree AV block: V=2/8; P=0 /9 Blood Pressure; Data was obtained from graphs in figure 1, page 902. The calculated weighted mean BP change was: Verapamil (n=8 ): SBP‐4.77±19 ; DBP‐4.4±6.18 Placebo (n=9 ): SBP‐7.5 ±16 ; DBP‐4±6.6 Standard deviation of change was not reported but imputed from end point Heart rate Data was obtained from graphs in figure 1, page 902. The calculated weighted mean HR change was: Verapamil (n=8 ): ‐4±23.8 Placebo (n=9 ): ‐9±13.32 Standard deviation of change was not reported but imputed from end point
Notes	Funding: British Heart Foundation and by MRC Programme Grant PG
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomized" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	High risk	"single‐blind" comment: double‐blind trials are considered yes to blinding
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Data reported according to methods

Di Pasquale 1994.

Methods	Single‐site study (Italy) Single‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 3 days (comparison), then same active treatment in both study groups after discharge Follow‐up: 6 months
Participants	371 patients within 4 hours after onset of symptoms of AMI Inclusion criteria: First AMI*, KK I‐II, acceptable echocardiography. * Diagnosis of AMI was based on ST elevation 1 mm in peripheral, 2 mm in precordial, blood CK‐MB within normal range. Exclusion criteria: Not suitable for thrombolysis, Left blundle branch block, cardiomyopathy, CHF, not satisfying reperfusion (relief of pain, peak of CPK within 12 h, early ventricular arrhythmias) unstable angina (not elevation of CPK ) Attrition data: Screened: 479, excluded at screening: 108 (78 not suitable for thrombolysis, 16 with KK > 3‐4, 14 with CK greater than normal) Total randomized patients: 371, Captopril : n= 188, No treatment : n= 183 Total withdrawals (discontinuation of drug): 112, Captopril: 57, No treatment: 55 Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: not reported based on all randomized patients
Interventions	Captopril   vs. No treatment  Drug regimen: Captopril : 6.25 mg orally  15 min before thrombolysis; 6.25 mg every 8 hours for 2 days, 12.5 mg every 8 hours for 4 days, followed by 25 mg  every 8 hours until after discharge. No treatment (control group): received no captopril for first 3 days, and then started captopril on the 4 day as above Other interventions: All patients received thrombolysis ( urokinase 1 million IU was administered) Captopril group: 58 patients received IV metoprolol No treatment  group: 49 patients received IV metoprolol
Outcomes	Mortality:  obtained from personal communication ( email annexed in appendix 2) Captopril: n= 188, at day 2: NR ; at day 10 ( in hospital ): 2;  at eof (6 mo ): not based on intention to treat analysis   No treatment (control group): n= 183, at day 2: NR; at day 10 (in hospital): 5; at eof (6 mo): not based on intention to treat analysis Total non‐fatal SAE: NR Blood Pressure: NR  Heart rate : NR
Notes	Funding: NR Dates of conducting the trial : Nov 30 1988 to Dec 31 1992 Author successfully contacted As confirmed by the author, there is another publication of the same trial (Int J cardiol: 46;107‐112) See appendix 2
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"Patients were randomized" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	High risk	Claimed to be single‐blind. Comment: we answer yes to blinding only with double‐blind designs
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	not reported
Incomplete outcome data addressed? 10‐day mortality	High risk	Out of 317, 112 were excluded from their analysis, we contacted trialist obtained additional information
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	not reported
Free of selective reporting?	High risk	There were two citations: . International Journal of Cardiology 1994; 43: 43‐50. and International journal of cardiology 1994; 46: 107‐112. The latter cited the former as if it was a different trial. There are differences in the number of patients "randomized" in those two trials. In the former trial, patients were randomized to captopril (n=188) or  no treatment (n=183) for 3 days. In contrast, in the latter trial, patients were randomized to captopril (n=84) or placebo (n=82) for 3 days. The author responded to us that these are two citations of the same trial. We are still not satisfied with explanation about the discrepancies.(see appendix)

Di Pasquale 1997.

Methods	Single‐site study (Italy) Double‐blind Method of randomization: NR Concealment of allocation: it was carried out by sequentially numbered boxes and was decided before thrombolysis Duration of treatment: 3 days (comparison), then same active treatment in both study groups for 10 days Follow‐up: 10 days
Participants	61 patients within 4 hours after onset of symptoms of suspected AMI Inclusion criteria: First episode of anterior AMI*, killip class I‐II, an acceptable echocardiographic window and admission to the hospital within 4 h of the onset of symptoms (pain). * AMI was base on ECG ST elevation of 1 mm in the peripheral leads and 2 mm in precordial leads,  with concomitant alteration of the segmentary kinetics in the echocardiogram performed at entry. The basal creatine kinase (CK, CK‐MB before thrombolysis) had to be within the normal range. All the patients admitted into the study had to have successful reperfusion. There was no age limit Exclusion criteria: Patients who were not suitable for thrombolysis, who had left bundle branch block on the admission ECG, a history of cardiomyopathy, or heart failure were excluded. Patients who did not satisfy the reperfusion criteria (first 4‐6 h after thrombolysis), and those receiving ACE inhibitors and [beta]‐blockers also were excluded from the study. Attrition data: Screened: 123, excluded at screening (31‐ for no or late thrombolysis, 5‐KK >3, 26‐inferior AMI) Total randomized patients: 61, captopril: n=31, Placebo : n= 30 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: captopril: 8/ 31 (26%) (4 patients with no reperfusion and 4 with unstable angina), Placebo: 8/30 (26%) (5 patients with no reperfusion and 3 with unstable angina) These patients were not considered for the analysis of efficacy, it is not known if those were accounted for safety Baseline characteristics: Not available for all randomized patients
Interventions	Captopril vs. Placebo Drug regimen: Captopril: 6.25 mg orally within 4 hours of starting thrombolysis, then up to 25 mg every 8 hour for 3 days. (The dose was increased (to maintain SBP > 100 mm Hg) Placebo: same regimen, No further details; although this group received captopril 72 hours after thrombolysis.  Other interventions: All patients received standard treatment: nitrates, heparin, aspirin, and where possible, metoprolol i.v., 5 mg, 3 times. The thrombolytic drug used was recombinant tissue‐type plasminogen activator (RTPA) (100 mg).
Outcomes	Mortality:  obtained from page 206 in results section, Captopril (n=31): at day 2 : NR; at day 10:2/31; at eof: NR  Placebo (n=30): at day 2: NR; at day 10:2/30; at eof: NR Total non‐fatal SAE: NR  Blood Pressure: NR  Heart rate: NR
Notes	Funding: NR Dates of conducting the trial: June 1994 to June 1995
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomized" no further details
Allocation concealment?	Low risk	it was carried out by sequentially numbered boxes
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	not reported
Incomplete outcome data addressed? 10‐day mortality	Unclear risk	Total lost to follow‐up: captopril: 8/ 31 (26%) (4 patients with no reperfusion and 4 with unstable angina), Placebo: 8/30 (26%) (5 patients with no reperfusion and 3 with unstable angina)
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	not reported
Free of selective reporting?	High risk	These patients were not considered for the analysis of efficacy, it is not known if those were accounted for safety

Durrer 1982.

Methods	Single‐centre (Netherlands), Open‐label Duration of treatment: 24 hours Follow‐up: 1 month Method of randomization: not reported Method of allocation: Consecutive order a series of sealed envelopes containing randomized allocations. Separate batches of different‐colored envelopes were used for early and late admissions. The envelopes were issued in balanced series of 20 to ensure that the number of patients in each group would be approximately equal. Each batch of envelopes was shuffled by two people to ensure proper randomization
Participants	328 patients within 4 or 12 hours from onset of AMI* Inclusion criteria: Patients with acute myocardial infarction criteria (*precordial pain that lasted for at least for one hour accompanied by sweating and pallor and resistant to sublingual nitroglycerine; ECG‐ Q waves or QR or QS complexes, ST changes) There were two subgroups: 1‐ admitted within four hours after the onset of symptoms of myocardial  infarction and; 2‐ admitted between 4 and 12 hours Exclusion criteria: If any doubt of AMI, cardiogenic shock  with SBP < 90 mm Hg, or urinary output <20 ml, pulmonary edema, SBP < 95 mm Hg, Heart rate >120 bpm, 2 or more previous MI Attrition data: Screened: 906, excluded at screening 578. Total randomized patients:328, nitroprusside (N): n= 163, and placebo n=165 Total withdrawals (discontinuation of drug): 7 (nitroprusside group) Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics:   Age(yr) N:60.9±11.4, P:61.4±10.9 SBP(mmHg): N: 137±22.2, P:136.6±21.2 Male/female: N: 124/39, P:131/34 Anterior localization: N: 86 (53%), C:88(54%) Medical history: Nitroprusside: previos AMI 34, CHF 15 Placebo: previos AMI 28, CHF 23
Interventions	Nitroprusside vs. Placebo Drug regimen: Nitroprusside: iv continuous infusion (24 hours) to bring SBP to a target of 100 mmHg; the infusion started at 15 mcg/min with increases of 5 mcg ‐ until maximum dose allowed of 500 mcg/min or a total of 720 mg. After 24 h infusion was tapered down and the drug was replace with oral isosorbide 5 mg  q 4 hours for the next six days. Total mean dose of nitroprusside was 183 ± 193 mg/h “(127 mcg /min)” Placebo: 5% glucose, no further details were given. It is not mentioned if patients took isosorbide after placebo Other interventions: All patients received standard medical treatment which was identical in both groups of patients: sublingual NTG, 2.5 mg of droperidol and 0.05 of fentanyl were used to relieve the pain. Beta blockers were used in 33% of nitroprusside group and 26% in placebo group.
Outcomes	Mortality: obtained from table 2 and text on page 1124 Nitroprusside: 2d:1/163 ; 10d:5/163  ; eof(1m.):9/163   Placebo:         2d:8/165 ; 10d: 18/165  ; eof(1m.):20/165 Total Non‐fatal SAE: NR Blood pressure: Data was obtained from table 1 and 2, page 1123, and was limited to one single point at hour 4 after starting treatment. The calculated mean change was: Nitroprusside:  SBP‐34.1± 12.2; DBP‐20.4±8.7 Placebo: SBP‐14.6± 18.9; DBP‐9.7±11.5 Standard deviation to change was not reported but imputed from end of treatment Hear rate: NR
Notes	Funding: The Wijnand M Pon Foundation. The trial used a particular method for closing‐out. Patients in the treatment and control groups were paired on the basis of their entry numbers‐ for example the fifth patient allocated to nitroprusside was paired with the fifth patient who received placebo. Then, the outcomes were scored according to those pairs as a tie, in favour of nitroprusside or placebo, respectively. However, they used a "closed" plan which meant that the trial had to end when a maximum of 62 pairs yielding preferences had been plotted. The critic of this is that it is not clear how the pairs were chosen, i.e., consecutively, randomly, or by outcome. If a total of 62 pairs were chosen and there were a total of 328 randomized patients, then it seems that the pairs were not chosen in consecutive order or randomly.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Consecutive order a series of sealed envelopes containing randomized allocations. Separate batches of different‐colored envelopes were used for early and late admissions. The envelopes were issued in balanced series of 20 to ensure that the number of patients in each group would be approximately equal. Each batch of envelopes was shuffled by two people to ensure proper randomization
Allocation concealment?	Unclear risk	Although it is stated that "consecutive order a series of sealed envelopes containing randomized allocations" were used there is another statement that potentially contradicts adequate concealment : "initial entry was arrange by the physician in charge... but in no case was necessary to reverse the original decision"
Blinding? All outcomes	High risk	There was not a statement for blinding
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	There is no report on lost to follow‐up or withdrawals
Free of selective reporting?	High risk	Early stop of the trial. The trialist used a particular method for closing‐out. Patients in the treatment and control groups were paired on the basis of their entry numbers‐ for example the fifth patient allocated to nitroprusside was paired with the fifth patient who received placebo. Then, the outcomes were scored according to those pairs as a tie, in favour of nitroprusside or placebo, respectively. However, they used a “closed” plan which meant that the trial had to end when a maximum of 62 pairs. They used a skew version of the plan with the lower boundary  much shorter than the upper since they “did not wish to determine whether nitroprusside might be worse than placebo” .

Eichler 1985.

Methods	Single‐site study (South Africa) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 36 hours Follow‐up: 3 days
Participants	32 patients with AMI* within 12 hours of onset of chest pain. Inclusion criteria: * History and ECG changes characteristic of AMI admitted to CCU within 12 hours from onset of chest pain. Exclusion criteria: Second or third degree AV block Sustain tachycardia SBP < 90 mm Hg Heart failure (wedge pressure > 25 mmHg Patients on current anti‐arrhythmic B‐blocker Digitalis Calcium antagonist therapy Attrition data: Total randomized patients: Tiapamil (T) group n= 16 Placebo (P) group n= 16. Total withdrawals (discontinuation of drug): 2 T: 1 P: 1 Withdrawals due to adverse events: 2 T: 1 (ventricular fibrillation) P: 1 (chest pain) Total lost of follow‐up: NR Baseline characteristics: Age(yr) T:55.1±10.5 P:51.6±11.6 SBP(mmHg) T: 128±22 P:135.6±24 Male/female T: 11/4 P:13/2 Anterior/ inferior localization T: 9/6 P:7/8 History of previous AMI: NR
Interventions	Tiapamil vs Placebo Dose regimen tiapamil as a load dose of 1 mg /kg over 3 min followed by an infusion of 25 mcg/kg/min over 36 hours. placebo (equivalent volumes of saline solution) Mean dose was not reported Other interventions: (number of patients) Tiapamil group: Atropine 1, furosemide 4, nitrates 1, fibrinolytic NR. Placebo group: Atropine 1, furosemide 3, nitrates 10 *, fibrinolytic NR.
Outcomes	Mortality: nil (text page 781) Total non‐fatal SAE: NR Blood Pressure: Data was obtained from table 2, page 783 The calculated weighted mean BP change was: tiapamil (n=16 ): SBP‐9.85 ±9 ; DBP‐12.85 ±13 Placebo (n=16): SBP 0.42 14± ; DBP 2.28 ±13 Standard deviation of change was not reported but imputed from end point Heart rate Data was obtained from table 2, page 783 The calculated weighted mean HR change was: tiapamil (n=16 ): ‐8.65±17.8 Placebo (n=16): ‐1.21±17.7 Standard deviation of change was not reported but imputed from end point
Notes	Funding: Hoffman La Roche (Switzerland) supplied drugs
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomly assigned" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Data reported according to methods

Erbel 1988.

Methods	Single‐site study (West Germany) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: hospital stay (between 3‐4 weeks) Follow‐up: hospital stay (between 3‐4 weeks)
Participants	149 patients within 6 hours after onset of symptoms of AMI Inclusion criteria: Diagnosis of AMI was based on acute chest pain lasting for more than 30 minutes and persistent ST segment elevation of more  than 0.3 mV in leasds V1‐V5 or 0.2 mV in leads I‐III, avl, avf,. Exclusion criteria: Long period of resusicitation, history of allergic reaction to streptokinase, previous cerebrovascular accident, surgery during the preceding 10 days, history of acute peptic ulcer, and history of bleeding. Attrition data: total randomized patients: nifedipine (N): n= 74, placebo (P): n= 75 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost of follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age: N:53±11, P:59±9 Gender (M/F): N:66/8, P: 59/16 Time to treatment (min): N:162±54, P:179±74 Medical history (number of patients) Nifedipine: previous AMI NR, diabetes 6, hyperlipidemia 19, smokers 45, Placebo: previous AMI NR, diabetes 14, hyperlipidemia 25, smokers 48
Interventions	Nifedipine vs. Placebo Drug regimen Nifedipine: Two 10 mg of nifedipine capsules sublingually in addition, nifedipine 0.2 mg intracoronarily was administrated before and after reperfusion of the vessel [after opacification of the infarct‐related vessel]. During their hospital stay patients received nifedipine 20 mg orally three times a day P: placebo; no further details. Other interventions: All patients received intravenous heparin 5000 IU (Overlapping  acetylsalicylate 1 gm and streptokinase 250,000 IU [ 20 minutes before catheterization] plus 250,000 IU intracoronarily for a total dose of 500,000 IU. After streptokinase, coronary angioplasty was attempted in patients with coronary lesion of more than 75%. Nifedipine group: IV nitroglicerine 36, oral isosorbide 28, beta‐blockers 15, diuretics 11, Placebo group: IV nitroglicerine 33, oral isosorbide 29, beta‐blockers 14, diuretics 12
Outcomes	Mortality:  obtained from fig 9, page 534 Nifedipine (n=74) : 2day  5; at 10 days: 10; eof: N/A Placebo (n=75) : 2day  1; at 10 days: 6; eof: N/A Total non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR Heart rate change during first 24 hours: NR
Notes	Funding: NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomly assigned", no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Data reported according to methods

ESPRIM 1994.

Methods	Multi‐centre (international‐117) Double‐blind Method of randomization: stratified in blocks of 8 and by study centre. Method of allocation: by a centralised computer system. Duration of treatment: 14 days  Follow‐up: 13 months
Participants	4,017 patients within 24 hours of the onset of a suspected AMI*, *based on typical persisting chest pain (>30) or less than 30 min but with definite CAD history. Exclusion Criteria: overt heart failure, definite indications or contraindications for intravenous vasodilators, pregnancy, or previously randomised in this RCT. Attrition data: screened: NR Total randomized patients: Molsidomine (nitrate‐ N): n= 2007, Placebo (P): n=2010 Total withdrawals (discontinuation of IV drug): N: 177/2007 (8.8%), P: 191/2010 (9.5%) Total withdrawals due to adverse events (IV drug): NR Hypotension: N: 1.9%, P:1.5% Cardiogenic shock: N: 0.9%, P:0.9% Death: N: 1.8%, P:2.4% Total lost of follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Male/Female: Molsidomine (nitrate‐ N): 1470/537, Placebo (P):1473/537 Age‐years: N=63.6±12.8, P=63.4±10 SBP: N=135±22, P=135±23 DBP: N=81±13, P=81±14 Medical history (%) Molsidomine, nitrate‐ (N) group:  previos AMI 14.9, previous angina 36.7, diabetes 13.4, hypertension 40.5. Placebo (P): previos AMI 15.7, previous angina 37.6, diabetes 12.3, hypertension 40.4.
Interventions	Molsidomine vs. Placebo Dose Regimen: Molisidomine: Intravenous linsidomine (molsidomine) 1 mg/ h for 48 hours (adjusted until from 0.2 mg to a maximum of 2 mg /h); followed by Oral therapy with molsidomine 4x4 mg tablets daily for 12 days Placebo: patients received placebo under exactly the same conditions, IV and oral therapy (no further details). Treatment started within 24 hours of onset of pain and lasted for 48 hours   after randomization. Mean duration of iv treatment : 44.8 h Other interventions: At investigator discretion : before randomisation about half of the patients had received thrombolytic  therapy, aspirin, and heparin , 30% received iv nitrates. After randomisation 86% received aspirin, 61 subcutaneous heparin, 74% iv heparin, and 20% thrombolytic therapy. More than 15% received nitrates.
Outcomes	Mortality: (day 2‐ text on page 93; day 10‐figure 1 page 93, month 13‐ text on page 94 ) Molsidomine: 2d:44/2007 ; 10d:127/2007 ; eof (13m): 294/2007  Placebo:        2d:58/2010;   10d:123/2010 ; eof (13m): 285/2010 Total Non‐fatal SAE: NR Blood pressure data: NR HR: NR
Notes	Funding:  Hechst (France) Cassella (Germany) Trial conducted on: June 1990‐ November 1992
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomization was done by computer system and was stratified in blocks of 8 and by study centre
Allocation concealment?	Low risk	Patients were assigned molsidomine or placebo treatment by a centralised computer system
Blinding? All outcomes	Low risk	double‐blind:"patients received placebo under exactly the same conditions as study drug"
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Free of selective reporting?	Low risk	Data reported according to methods

Fitzgerald 1990.

Methods	Multi‐centre (9 in England). Double‐blind Method of randomization: NR Method of allocation: NR Duration of treatment:  5 days Follow‐up: 6 months
Participants	360 patients within 8 hours of the onset of an AMI. Inclusion criteria: Patients thought with acute myocardial infarction in the past 8 hours. (*clinical findings, and systolic blood pressure  greater than 90 mm Hg Exclusion criteria: Had received long‐acting nitrates within 8 hours. Physician whishing to prescribe nitrates within 8 hours. Attrition data: Screened: NR, Total randomized patients: Isosorbide 5 mono‐nitrate (N ): n=184, Placebo (P): n= 176 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: 7 in isosorbide group and 4 in the placebo group Baseline characteristics: Age, years: Isosorbide 5 mono‐nitrate (N ): 60 [33‐74, Placebo (P): 60 [38‐74 BP (mmHg): N: NR, P: NR Site of infarct N: anterior 77, inferior 82, posterior 1, unknown 11, not MI 13 P: anterior 79, inferior 79, posterior 0, unknown 5, not MI 13 Medical history N: 1 st MI 138, 2 nd MI or more 22, unknown 24 P: 1 st MI 136, 2 nd MI or more 29, unknown 11
Interventions	Isosorbide 5 mono‐nitrate vs Placebo Drug regimen: Isosorbide, Initial: oral 40 mg if SBP was >110 mmHg or 20 mg if SBP ? 110  mm Hg. Maintenance: oral 20 mg qid x 48 hours, 20 tid x 24 hours and 20 bid x 48 hours (total duration of treatment 5 days). Medication was withdeld if SBP was < 90 mm Hg or if HR was outside the 50‐130 range. Medication was re‐commenced if patient recuperated the allowed range. Placebo: no details were given. Other interventions: Lignocaine: N: 23%, P: 21% IV nitrates: N: 15%, P: 25% Diamorphine: N: 42%, P: 36%
Outcomes	Mortality: (text [in percentage], on page 122) Isosorbide: 2d:NR ; 10d: 9/184  ; eof (6 m): 26/184  Placebo:     2d: NR ; 10d:7/176  ; eof(6 m): 19/176 Total Non‐fatal SAE: NR Blood Pressure; Data was obtained from graphs fig 1& 2, on page 122. The calculated weighted mean BP change (at 24 h) was: Isosorbide (n=184): SBP‐19.25±29; DBP‐8.75 ±17.56 Placebo (n=176): SBP‐13.05±13.26; DBP‐6.68±17.14 Standard deviation of change was not reported but imputed from end point Heart rate Data was obtained from graphs fig 3, on page 122. The calculated weighted mean HR (at 24 h) change was: Isosorbide (n=184): ‐0.22±12.76 Placebo (n=176): 3.25±12.48 Standard deviation of change was not reported but imputed from end point
Notes	Funding: Schwarz Pharma Limited
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"This study was randomized, double‐blind", no further details. Probably done adequately
Allocation concealment?	Unclear risk	not reported, no comment
Blinding? All outcomes	Low risk	"This study was randomized, double‐blind", no further details. Probably done adequately
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Low risk	Total lost to follow‐up: 7 in isosorbide group and 4 in the placebo group
Free of selective reporting?	Low risk	Mortality data reported according to methods

Flaherty 1983.

Methods	Single‐centre (US). Single‐blind Method of randomization: A card was drew to randomized patients Concealment of allocation: NR Duration of treatment: 48 hours Follow‐up: 15 months (range 4‐44)
Participants	104 patients with AMI within 12 hours from onset of chest pain. Inclusion criteria: Patients with high probability of acute myocardial infarction (history and electrocardiographic changes) within 12 h from onset of chest pain. Exclusion criteria: Age >75 h SBP < 90 mm Hg Second or third degree AV block Heart rate: < 55 beats/min Severe hypertension (SBP>200 mmHg, DBP>120) after relief of pain Attrition data: Screened: NR, Total randomized patients: 104, Nitroglycerin (N): n=56, Placebo (P): n= 48 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up:  3 patients at month 3 Baseline characteristics for the two randomized groups:  (mean values ± SD) Age (years) Nitroglycerin (N): 60±11, Placebo (P): 57±9 SBP (mmHg): N: 142±27, P:150±29 DBP (mmHg): N: 90±18, P:96±18 Male (%): N: 64, P:69 Anterior/ transmural localization (%): N: 44/84, P: 36/77 Medical history (%) N: previous infarction 16,previous heart failure 38 P: previous infarction 27,previous heart failure 34
Interventions	Nitroglycerin vs. Placebo Dose regimen: Nitroglycerin: intravenous initial dose was begun at 5 mcg /min with stepwise increases at 3‐5 min intervals, followed by an adjustment phase to lower MAP 10 % from baseline. The infusion continued x 48 hours Placebo :  equivalent volumes 20‐30 ml/h of 0.9 ml of 95% ethanol added to 500 ml of 5% dextrose in water to match the composition of the vehicle in which  nitrolgycerin was dissolved. The infusions were terminated abruptly after 48 hours . Then patients received nitroglycerin ointment 2%, or placebo ointment every 4 hr for 72 h, to maintain BP  levels previously attained with iv infusion. The mean dose of NTG to obtain 10% lowering of MAP was 90±74 mcg/min The mean duration of the Nitroglycerin titration period was 81±63 min (range 9 to 425) The mean duration of placebo titration was 27±23 min (range 0 to 115). Other interventions: Patients with Killip class I‐II received lidocaine 1 mg/kg bolus and 20 mcg/kg/min for at least 24 hours. Patients with Killip class III‐ received half of the above dose. Morphine, sublingual NTG, diuretics, digitalis and Antihypertensive were allowed. Heparin by constant infusion was given to all patients. Propranolol in case of a recurrent ischemia. The actual number of patients who receive the above co‐interventions was not reported. Thrombolysis was no mentioned
Outcomes	Mortality: (based on text ?page 580) Nitroglycerin: 2d: NR ; 10d:4/56  ; eof(3 m.):11/56  Placebo:        2d:NR ; 10d: 7/48 ; eof(3 m.):11/48 Total non‐fatal SAE: NR Total individual SAE: Blood pressure: Data obtained from text, page 579. The weighted mean BP change was calculated from a single point after baseline: Nitroglycerin (n=56): SBP‐19 ±21; DBP‐9 ±15 Placebo (n=48): SBP‐4 ±29 ; DBP‐3 ± 19 Standard deviation of change was not reported but imputed from end point Heart rate Data obtained from text, page 579. The weighted mean HR change was calculated from a single point after baseline: Nitroglycerin (n=56):‐2 ±15 Placebo (n=):‐2 ±15 Standard deviation of change was not reported but imputed from end point
Notes	Funding:  National Institute of Health (NIH) Dates of conducting the trial : NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	A card was drew to randomized patients. No further details
Allocation concealment?	Unclear risk	Not reported, no comment
Blinding? All outcomes	High risk	"patients were entered into this randomized, single‐blind trial"; no further details Comment: we only considered double‐blind trials as blinding
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Free of selective reporting?	Low risk	Mortality data reported according to methods

Galcera 1993.

Methods	Single‐site study (Spain) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 14 days* Follow‐up: 14 days * The exact duration of treatment is not reported. It is assumed that it is 14 days as the primary objective of the trial was to determine short‐term remodelling (by radionuclide ventriculograms at baseline and at weeks)
Participants	43 patients within 24 hours after onset of symptoms of first AMI. Inclusion criteria: ST segment elevation in at least 3 precordial leads (anterior infarction), with retrospective documented development of Q waves and typical changes in enzymes levels. Under 70 years, availability for radionuclide ventriculography within the first 24 h.   Exclusion criteria: Patient refusal to participate, previous valvulaopathy, Kiillip class III or IV, and enzymatic evidence of infarct extension Attrition data: Screened: 85, excluded at screening 13 ( 2 died on admission, 2 hemodynamic study not available, 9 not available for radionuclide); other 29 patients who had pulmonary capillary wedge pressure less than 17 mm Hg  were not randomized (the latter point was not pre‐specified in the inclusion criteria) Total randomized patients: Captopril: n= 22, Placebo : n= 21 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: NR based on all randomized patients
Interventions	Captopril  vs. Placebo Drug regimen: Captopril : Oral 6.25 mg titrated to a target dose of 25 mg tid Placebo: No further details Other interventions: All patients received conventional therapy including intravenous nitroglycerin, patients could received CCB, BB, diuretics digitalis at physician discreation. Thrombolysis: Captopril: 15 / 22 (68%), Placebo : 16 / 21(76%)
Outcomes	Mortality:  obtained from text , page 260 Captopril : 2day:NR ; 10day (before day 14): 1/21 ; eof(14 d ): NR Placebo : 2day:NR ; 10day (before day 14): 2/22 ; eof(14 d ): NR Total non‐fatal SAE: NR   Blood Pressure change; NR Heart rate change: NR
Notes	Funding: NR Dates of conducting the trial: February 1988‐ June 1990
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"this study was a randomized.." no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	High risk	Base‐line variables not reported based on all randomized patients

Gelmers 1988.

Methods	Multi‐centre (4, Netherlands ) Double‐blind Method of randomization: Patients were randomly assigned (in block of six) Concealment of allocation: NR Duration of treatment: 28 days Follow‐up: 6 months
Participants	186 patients within 24 hours after onset of symptoms of stroke Inclusion criteria: Patients over 45 years, completed acute ischemic stroke, admission to hospital within 24 after the onset of symptoms. *Diagnosis of stroke was based on the sudden onset of persistent focal neurologic deficit with no subsequent progressive deterioration. It was confirmed by a complete neurologic workup that included a CT scan. Exclusion criteria: All patients with brain ischemia caused by factors other than atherothrombosis, such as subarachnoid haemorrhage, intracerebral haemorrhage, haematoma, and complicated migrane. Patients with overt, severe systemic disease ( myocardial infarction, cardiogenic shock, severe renal or hepatic failure, severe systemic infection, non‐stabilized mellitus). Attrition data: Screened and excluded after screened: NR Total randomized patients: 186, Nimodipine: 93, Placebo : 93 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost of follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years): Nimodipine (N):69.3 ±11.9, Placebo (P): 69.6 ±10.7 Gender: N: female 44, male 49 , P: female 31, male 62 Localization of the stroke: N: left 55, right 38 ; P: left 61, right 32 Medical history: N: hypertension 21, diabetes 16, cardiac disorders 32 P: hypertension 21, diabetes 16, cardiac disorders 31
Interventions	Nimodipine vs. Placebo Drug regimen: Nimodipine: 30 mg every 6 hours for 28 days Placebo: tablets identical in appearance and taste to the drug and according to the same schedule. Other interventions: All patients received standard therapy of 10 percent depolymerised dextran of low molecular weight for 12 hours a day for five days. Low dose heparin sodium (10,000 IU in two divided doses) for prevention of deep venous thrombosis. During the study, patients were not allowed to use vasodilators, steroids, antiplatelet agents, barbiturates, or hyperosmolar agents. Cardiotonic agents, anti‐hypertensive medications and antibiotic were administrated as required. No calcium‐channel blockers other than nimodipine were administrated
Outcomes	Mortality:  obtained from graph figure 1, page 205 Nimodipine: 2day: NR; 10day (7day):4/93; eof (4w‐N/A):8/93  Placebo: 2day: NR; 10day (7day):9/93; eof (4w‐N/A ):19/93 Total non‐fatal SAE: NR Blood Pressure data: NR Heart rate data: NR
Notes	Funding: Unrestricted grant from Bayer AG, Wuppertal, Germany Dates of conducting the trial: September 1982‐1984
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Patients were randomly assigned (in block of six)
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

GISSI‐3 1994.

Methods	Multi‐centre (200 centres, 2/3 Italy). Open‐label 2 x 2 factorial design Method of randomization: computer network system Method of allocation: Program based on biased coin algorithm Duration of treatment: 6 weeks Follow‐up: 6 months
Participants	19,318 patients within 24 h of onset of suspected AMI Note; the diagnosis of AMI was confirmed in 95% of randomized patients, acute coronary syndrome in 3.6%, other diagnosis in 1.4% Inclusion criteria: Patients with chest pain and ST elevation of 1 mm in peripheral leads, or 2 mm in precordial leads, within 24 hours of symptoms onset and having no contraindications. Exclusion criteria: Severe heart failure, Killip class 4; high risk of adverse effects, contraindications such as renal failure, bilateral renal artery stenosis, allergies, life threatening disorders, previous randomization. Attrition data: Screened: 43,047,  excluded at screening: 23,653: 33.2%‐ admitted after 24 hours 23.2% ‐contraindication to study drug 15.4%‐ persistently unstable 28.2%‐had administrative reasons Total randomized patients: 19,394 (76 (0.39%) were lost to follow‐up, thus 19,318 were included in the mortality analysis; but 499 patients were not included in the original paper leaving a total 18,895, in 1994) Glyceryl trinitrate: n= 9,663 (9,453 included in the 1994‐original paper) Control ‐ no glyceryl trinitrate : n= 9,655 (9,442 included in the 1994‐original paper) Lisinopril : n= 9,646 (9,435 included in the 1994‐original paper)  Control ‐ no Lisinopril : n=9,672 (9,460 included in the 1994‐original paper) Total withdrawals (discontinuation of drug): 2,745 /18,895 (14.5 %) Glyceryl trinitrate : 1089/ 9453 (11.5%) Lisinopril : 1656/ 9,435 (17.56%)   Total withdrawals due to adverse events: NR Glyceryl trinitrate : hypotension 2.6 %, headache 3.0 % Lisinopril : hypotension (9.7%), renal function impairment (2 %), coughing (0.5 %).   Total lost to follow‐up: (from total randomized 19,394) Up to 6 weeks: 76 (0.39%) [Reported in J Am Coll Cardiol 1996; 27:337‐344] the distribution according to groups is not reported. Baseline characteristics: Abbreviation: Glyceryl trinitrate (N ), Control group ‐no glyceryl trinitrate (XN), Lisinopril (L), control group‐ no Lisinopril (XL) Women (%): L: 22.3, XL: 22.1, N: 22.4 , XN: 22 Age > 70 yr : L: 26.8, XL: 27.4, N: 26.9, XN: 27.3 Medical history (%): previous AMI, angina, treated hypertension, diabetes L: 14, 34.4, 30.2, 15.5.  XL: 13.8, 34.4, 29.6, 15.7. N:  13.7, 34.5, 30.6, 15.9. XN: 14.1, 34.5, 29.2, 15.4. Time from symptom onset:  <6;  >6‐12; >12‐24 hours L:  34.9; 24.6; 40.5 XL: 34.8; 25.2; 40 N: 34.8; 25.3; 39.9 XN: 34.9; 24.5; 40.6
Interventions	Glyceryl trinitrate vs. open control and Lisinopril vs. open control   Drug regimens Glyceryl trinitrate was initially started as IV infusion at a rate of 5 mcg x min with adjustments every 5 minutes of 5‐20 mcg/min to achieve a target of systolic blood pressure reduction of at least 10 % but not lower than 90 mm Hg, during the first 24 hours. After that the IV infusion was replaced with a patch providing 10 mg of nitrate transdermally per day. The patch was removed at bedtime. The duration of study treatment was for 6 weeks.  Note: if the patch was not tolerated a single oral dose of 50 mg isosorbide mononitrate was given daily. Control group ( open control) was given no study treatment ( no further details) Lisinopril was also given for 6 weeks. Initially, 5 mg every 24 hours was given. After 48 hours ,10 mg/ day was given until completion of 6  weeks. If systolic blood pressure was reduced (within the first 3 days) to less than 120 mm Hg a dose of 2.5 mg was given. If, at any time. systolic blood pressure was reduced to less or equal to 100 mm Hg, a maintenance dose of 5 mg daily could be adopted. If SBP was reduced to less than 90 the therapy was stopped. Control group ( open control) was given no study treatment ( no further details)   Other interventions: Lisnopril group: IV beta‐blockers 30.1% , fibrinolytics 71.4% , aspirin 83.6%, other antiplatelets 3.6% . Control‐ no Lisinopril group: IV beta‐blockers 31.3% , fibrinolytics 71.9% , aspirin 84.2% , other antiplatelets 3.5%. non‐study ACE‐I 13.3% Nitrate group: IV beta‐blockers 30.4%, fibrinolytics 71.9% , aspirin 84.1%, other antiplatelets 3.4%. non‐study nitrate not reported. Control‐ no nitrate group: IV beta‐blockers 30.9%, fibrinolytics 71.4%, aspirin 83.6, other antiplatelets 3.7%, non‐study nitrate 57.1% (mainly for angina, reported in page 1119 in Lancet 1994;343)
Outcomes	Mortality:  (data, at day 2 and 10, was obtained from personal communication‐see email Appendix 2) Lisinopril :            2d:200 ; 10d : 435/ 9,646; end of treatment (42 days):N/A 619/9646 Control lisinopril: 2d: 234 ; 10d: 486 / 9,672; end of treatment (42 days): N/A 693/9672   Nitrate:                 2d:196 ; 10d :443 / 9,663; end of treatment (42days): N/A 639/9663 Control nitrate:     2d:238 ; 10d :478 / 9,655; end of treatment (42days): N/A 673/9655 Total non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR Heart rate change during first 24 hours: NR
Notes	Funding: Zeneca Pharmaceutical (who supplied Lisinopril and Schwarz Pharma (who supplied intravenous and transdermal Glyceryl trinitrate). Dates of conducting the trial: June 1991 and July 1993.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Central Computer network system
Allocation concealment?	Low risk	Program  based on biased coin algorithm
Blinding? All outcomes	High risk	open‐label trial
Incomplete outcome data addressed? 2‐day mortality	Low risk	Not reported in the original trial but in personal communication
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients.
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Hargreaves 1992.

Methods	Single‐site study (UK) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 28 days Follow‐up: 28 days
Participants	105 patients within 24 hours after onset of symptoms of AMI Inclusion criteria: Suspected acute myocardial infarction within 24 hours of the start of chest pain and SBP > 90 mm Hg. Exclusion criteria: Not reported Attrition data: Screened: NR Total randomized patients: 105, Captopril: n= 36, Isosorbide mononitrate: n=33, Placebo : n= 36 Total withdrawals (discontinuation of drug): 18 but distribution according to group not reported Total withdrawals due to adverse events:  NR Total lost to follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years): Captopril (C): 60.6±9.4, Isosorbide mononitrate (N): 60.6±11.2, Placebo (P): 60.8 ±8.4 Sex ( Men/women ): C: 30/6 ; N: 30/3 ; P: 31/5 Site of infarction (anterior / inferior): C: 12/23, N: 15/18, P: 14/21 Previous infarction: C: 0(0%), N: 4(12%) , P: 4(11%) BP (mm Hg): NR
Interventions	Captopril   vs. Isosorbide mononitrate vs. Placebo Drug regimen: Captopril: orally, initial 6.25 mg, then, 12.5 mg  t.i.d. for 28 days Isosorbide mononitrate: 20 mg , no further details Placebo: no further details Other interventions: Thrombolysis Captopril: 32 (88%)   Isosorbide mononitrate: 30(91%) Placebo: 30 (83%)
Outcomes	Mortality:  obtained from text, page 370 Captopril: 2day:NR ; 10day: 0/36  ; eof(N/A, 28 days ):  NR Isosorbide: 2day:NR ; 10day: NR  ; eof(N/A, 28 days ): 2/33 Placebo: 2day:NR ; 10day: 2/36  ; eof(N/A, 28 days ):  NR Total non‐fatal SAE: NR Blood Pressure change during first 24 h: NR Heart rate change during first 24 h: NR
Notes	Funding: Bristol‐Myer Squibb and Stuart Pharmaceutical Date of conducting the trial : NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomised", no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind" trial, no further details. Probably it is done properly
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	High risk	dose regimen for isosorbide group was not reported

Heber 1987.

Methods	Single‐site study (UK) Open‐label Method of randomization: NR Concealment of allocation: NR Duration of treatment: 5 day Follow‐up: 1 year
Participants	166 patients within 6 hours after onset of symptoms suggesting AMI   Inclusion criteria: Under 75 years of age, diagnosis of AMI was based on clinical ECG and CK‐MB more than 15IU.   Exclusion criteria: Presentation more than 6 hours , left ventricular failure, persisting hypotension (SBP < 100 mmHg) or hypertension (SBP > 200 mmHg), or conduction disorders, tachyarrhythmias requiring treatment, haemodynamically valvular regurgitation, history of bronchospasm, hepatic or renal disease, recent treatment with verapamil. Attrition data: Screened: 630, total randomized patients: 166, labetalol: n=   83 No treatment: n= 83, Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR   Baseline characteristics: There were 137 men, 29 women, age 39‐74 years. Age (years): Labetalol (L): 61, No treatment (X): 58 Male/ female: L: 65/18, X: 72/ 11 SBP /DBP (mm Hg) : Labetalol: 144/83 , No treatment: 141/91 BP standard deviation at baseline: NR Medical history (%) Labetalol group: previous ischaemia 27(32.5), hypertension 18 (21.6), diabetes 3(3.6), No treatment group: previous ischaemia 33 (39.7), hypertension 9(10.84), diabetes 4 (4.81),
Interventions	Labetalol vs. No treatment Drug regimen: Labetalol: 0.25‐1 mg/ kg (according to SBP at baseline) initial IV bolus over 10 minutes, then 6 hour IV infusion 10 mcg /min. Then, oral treatment of 50‐200 mg every 8 hours was started for five days. The aim of was to maintain systolic blood pressure near but not below 100 mm Hg. The mean dose during IV infusion of labetalol was 23 ± 14 mg (range 2 ? 78 mg)  Other interventions: All patients were treated otherwise in a conventional manner: analgesics, diuretics, digoxin, atropine, antiarrhythmics, and inotropics. Nitrates and CCB were discourage during the acute phase. Labetalol group: furosemide 18, thiazides 24, atropine 5, digoxin 5, lignocaine 5, inotropes 2. No study treatment group: furosemide 14, thiazides 26, atropine 10, digoxin 4, lignocaine 4, inotropes 1.
Outcomes	Mortality:  obtained from text, page 16 Labetalol: 2day: 5/83; 10day (in hospital):5/83; eof( one year): 12/83  Placebo: 2day: 1/83; 10day (in hospital): 1/83; eof (one year ): 7/83 Note: 6 patients died within the first six hours (text, page 14)   Total non‐fatal SAE: NR Blood Pressure: Data was obtained from figure 1, page 14. The mean BP change at 6 h was: Labetalol (n= 83): SBP ‐30.9 ±NR; DBP ‐7.13 ± NR Placebo (n= 83): SBP ‐13.48± NR; DBP ‐7.48 ± NR Standard deviation was not reported Heart rate: Data was obtained from figure 2, page 15. The mean HR change at 6 h was: Labetalol (n= 83): ‐14.00 ± NR Placebo (n= 83 ): ‐5.12 ± NR Standard deviation was not reported
Notes	Funding: NR Dates of conducting the trial: February 1982 ‐September 1983
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomized" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	High risk	open control
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Free of selective reporting?	Low risk	Mortality data reported according to methods

Hildebrandt 1992.

Methods	Single‐site study (Denmark) Double blind: Method of randomization: blocks of four by computer Concealment of allocation: consecutively numbered treatment packages Duration of treatment: 48 hours Follow‐up: until hospital discharge
Participants	100 patients with suspected myocardial infarction (MI)* within 8 hours of onset. * MI:severe chest pain and ECG ST segment elevation of depression. Exclusion criteria: Cardiogenic shock, known bleeding disorders, contraindications to streptokinase; intolerance to study drugs, pregnancy, lactation. Attrition data: Total randomized patients: Isosorbide dinitrate ( N ): n= 50 Placebo (P): n= 50 Total withdrawals (discontinuation of drug): % N: 9 P: 16 Withdrawals due to adverse events: NR Total lost of follow‐up: 1 (excluded before initiation of treatment because of cardiogenic shock) Baseline characteristics: Mean age (years) Isosorbide dinitrate ( N ):64, Placebo (P):62 Male: (%): N: 78, P: 65 Blood pressure: NR Medical History: (%) Isosorbide: previous AMI 26, angina 36, hypertension 20, diabetes 6, hypercholesterolemia 12. Placebo: previous AMI 22, angina 31, hypertension 12, diabetes 12, hypercholesterolemia 6
Interventions	Isosorbide dinitrate ( N ): n= 50 Placebo (P): n= 49 Dose regimen: N: Intravenous infusion of isosorbide dinitrate at initial dose of 33 mcg /min to achieve 10 % reduction in SBP. Duration of treatment 48 hours or until adverse effect. P: identical 50 ml bottles at same rate Mean dose of isosorbide administrated was not reported 90% of patients completed 48 h of infusion Co‐interventions: previous to isosorbide patients received 1.5 million I.U. of streptokinase as bolus and 2 mg/h as infusion. Aspirin 150 mg orally, was also given. Diuretics and digoxin were given at the discretion of the staff.
Outcomes	Obtained from this trial for the two randomized groups: Isosorbide dinitrate ( N ): n= 50 Placebo (P): n= 49 Mortality: (from text on page 1141) Isosorbide: 2d:0/50; 10d:2/50. eof (1m): NR Placebo: 2d:2/49; 10d:6/49.eof(1m): NR Non‐fatal SAE: NR Any individual SAE: CHF: Isosorbide: 2d:NR ; 10d:14/50 ; eof(3 m.): NR Placebo: 2d:NR ; 10d:20/49 ; eof(3 m.): NR SHOCK: Isosorbide: 2d:NR; 10d:2/50 ; eof(3 m.): NR Placebo: 2d:NR; 10d:5/49 ; eof(3 m.):NR Blood Pressure: SBP: NR DBP: NR Heart rate : NR
Notes	Funding: Schwarz‐Pharma AG
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	in blocks of four by computer
Allocation concealment?	Low risk	consecutively numbered treatment packages
Blinding? All outcomes	Low risk	double‐blind trial; "dentical 50 ml bottles at same rate... as study drug"
Incomplete outcome data addressed? 2‐day mortality	Low risk	1 (excluded before initiation of treatment because of cardiogenic shock)
Incomplete outcome data addressed? 10‐day mortality	Low risk	1 (excluded before initiation of treatment because of cardiogenic shock)
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	well conducted trial

ICSG 1984.

Methods	Multi‐centre (8 countries ‐ Europe ) Double‐blind Method of randomization: randomization was performed within each center and was balance in blocks of four patients. Concealment of allocation: (double blind) Duration of treatment: during hospital stay Follow‐up: during hospital stay
Participants	144 patients within 4 hours after onset of suspected AMI   Inclusion criteria: Patients of either sex between 21 and 70 years of age; within four hours of the onset of symptoms of a suspected  first myocardial infarction*.   * patients were classified as A: if they had chest pain > 30 minutes by without ECG abnormality; or B: if they had ST‐elevation or depression of more than 0.5mm in standard inferiour leads, or more than 1 mm in chest leads V3‐V6 or new T‐wave inversion in V3‐V6 or II, III, and AVF.   Exclusion criteria: Bradycardia (<50 beats per minute); hypotension (SBP < 100 mm Hg), clinical evidence of severe left ventricular failure, any degree of heart block, or a history of previous myocardial infarction and a QRS duration longer than 0.11 second; contraindication to beta‐blockade ( rales > 10 cm above the diaphragm), bronchial obstruction; current treatment with a beta‐blockcker, CCB, digitalis or another antiarrhythmic agent.     Attrition data: Screened: NR, Total randomized patients: 144, Timolol: n=73, Placebo : n=71 Total withdrawals (discontinuation of drug): NR Total withdrawals due to adverse events: 24, Timolol (T): 14, Placebo (P): 10 Total lost to follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years): Timolol (T): 57±19 (32‐74), Placebo (P): 54±10 (33‐72) Gender : male/ female: T:62/11, P:61/10 Time to treatment (hours): T:3.4, P: 3.6 Infarct location: anterior; inferior; indeterminate Timolol (T):34, 22, 5 Placebo (P):34, 28, 1 SBP (mm Hg): Timolol (T):156/90, Placebo (P): 156/93 Neither DBP nor standard deviation of SBP at baseline were reported
Interventions	Timolol vs. Placebo Drug regimen: Timolol: Therapy had to be initiated within 5 hours of the onset. Treatment began with a bolus injection of 1 mg timolol maleate and repeated 10 minutes later (if hemodynamic stable), and then 0.6 mg /hour infusion for 24 hours. Then oral therapy 10 mg bid during hospital stay  Placebo: normal saline same dose regiment. No further details Other interventions: number of patients (milligrams of drug) Timolol group:   digitalis 6 (NR), diuretics 33 (furosemide 1280), analgesics 21 (morphine 630), beta‐blockers 3 (NR), nitrates 7(NR). Placebo group: digitalis 6 (NR), diuretics 37 (furosemide 1260), analgesics 33* (morphine 838*), beta‐blockers 8 (NR), nitrates 11(NR). * p< 0.05
Outcomes	Mortality:  obtained from text, page 11 Timolol: 2day:1/73 ; 10day (hospital stay):3/73; eof( N/A):  Placebo: 2day: 3/71; 10day (hospital stay):4/71; eof( N/A): Total non‐fatal SAE: NR Blood Pressure: Data was obtained from figure 1, page 11. The mean BP at end point (24 h) was: Timolol (n=73 ): SBP 119±NR ; DBP 69.94±NR Placebo (n= 71): SBP 132.41 ± NR ; DBP 78.26 ±NR Heart rate: Data was obtained from figure 1, page 11. The mean HR at end point ( 24 h) was: Timolol (n=73 ): 57.58±NR Placebo (n=71): 69.80 ±NR
Notes	Funding: Merck, Sharp and Dohme Research Laboratories, Rahway NJ Dates of conducting the trial : NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	randomization was performed within each centre and was balance in blocks of four patients
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods

Infeld 1999.

Methods	Multi‐centre (Australia) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 14 days Follow‐up:  3 months
Participants	50 patients within 12 hours after onset of symptoms of stroke Inclusion criteria: Acute middle cerebral artery territory cortical infarction* and if it was possible to obtain acute CT and SPECT scans, commence of treatment within 12 hours of stroke onset. * Diagnosis was made on the basis of either 1. presence of cortical neurological deficits such as dysphasia,anosognosia, visual or sensory inattention, dyspraxia, or parietal sensory deficit or 2. evidence of cortical infarction on the acute CT scan Exclusion criteria: Presence of cerebral hemorrhage or noncerebrovascular pathology such as tumor , previous cerebral pathology, presence of other neurological systemic or psychiatric illness, concurrent use of other dihydropyridine, 1 or more contrainidications  to CCB ( pregnancy, post‐partum, hpatic, renal or cardiac disease, or drugs affecting hepatic metabolism) Attrition data: screened/ excluded during screened: NR Total randomized patients: Nimodipine: n= 25, Placebo: n=25 Total withdrawals (discontinuation of drug): Nimodipine: 3 (died)/ 25, Placebo : 3 (died) / 25 Withdrawals due to adverse events: Nimodipine: 3 (died)/ 25, Placebo : 3 (died) / 25 Total lost to follow‐up: 4(2 in each group) were excluded for any analysis. Baseline characteristics: Female/ male: 26 men, 24 women Age (years):Nimodipine (N): 69.8 ± 2.5, Placebo (P):  70.7± 2.3 Treatment delayed (hours) N: 8.2 ± 0.6, P:  8.7 ± 0.5 SBP / DBP (mm Hg) : NR
Interventions	Nimodipine vs. Placebo   Drug regimen: Nimodipine: 120 mg/ day orally in 4 divided doses ( 30 mg every 6 hours)  x 14 days Placebo:  similar in appearance (No further details) x 14 days
Outcomes	Mortality:  obtained from text, page 1418 and 1419 Nimodipine: 2day: NR ; 10day (7 day):3/25 ; eof(3 mo‐N/A): NR Placebo: 2day: NR ; 10day (7 day):3/25 ; eof(3 mo‐N/A): NR Total non‐fatal SAE: NR Blood Pressure data : NR Heart rate data : NR
Notes	Funding: National Health and Medical Research Council and the National Stroke Foundation. The investigators are grateful to Bayer Australia Limited for supplying nimodipine and placebo Dates of conducting the trial : November 1993‐ May 1996
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomized" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind trial"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

INWEST 1994.

Methods	Multi‐centre (34‐Europe ) Double‐blind Method of randomization: predetermined randomization list Concealment of allocation: NR Duration of treatment: 21 days Follow‐up: 6 months (24 weeks)
Participants	295 patients within 24 hours after onset of symptoms of stroke Inclusion criteria: Clinical diagnosis of a recent (within 24 h) ischaemic stroke in the carotid artery territory , 40 year or older and functionally independent  before stroke, should be conscious with a stable marked hemiparesis, Mathew score of 65 or less or an Orgogozo score between 5 and 50. Exclusion criteria: Recent or unstable cardiac disease, any disorder interfering with the neurological or functional assessment or other life‐threatening concurrent illness. Attrition data: Total screened: NR Total randomized patients: 295 (7/295 other than ischaemic stroke) Nimodipine 1 mg: n= 101   Nimodipine 2 mg : n= 94 Placebo : n= 100 Total withdrawals (discontinuation of drug): 101 Nimodipine 1 mg: n= 36/101   Nimodipine 2 mg : n=38/ 94 Placebo : n= 27/100 Withdrawals due to adverse events: 16 (not reporting according to group) Total lost of follow‐up: 7 Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years):Nimodipine 1 mg (N1): 71.9, Nimodipine 2 mg (N2):  72.1, Placebo (P): 71 Male sex (%) N1: 45, N2: 45, P: 49 Previous vascular events (%) N1: TIA 12; RIND 2; stroke 17; myocardial infarction 9; atrial fibrillation 34 N2: TIA 12; RIND 2; stroke 13; myocardial infarction 13; atrial fibrillation 27 P: TIA 17; RIND 6; stroke 9; myocardial infarction 16; atrial fibrillation 32 SBP / DBP (mm Hg) N1: 158.6/87.1, N2: 161/89.4, P: 159.5/91.2
Interventions	Nimodipine vs. Placebo Drug regimen: Nimodipine: subdivided in two groups: 1 mg group and 2 mg group: received 1 or 2 mg / hour intravenous infusion for 5 days and then both groups receive 30 mg nimodipine orally 4 times a day (120 mg/day) for 16 days ( total treatment 21 days)   Placebo: No further details Other interventions: NR
Outcomes	Mortality:  obtained from text, page 207, and table 3 page 208 Nimodipine: 2day:NR; 10day (5day):25/195  ; eof( 6 months‐N/A):83/195  Placebo: 2day:NR ; 10day (5day):11/100  ; eof( 6 months‐N/A): 33/100 Total non‐fatal SAE: NR Blood Pressure; Data was obtained from graph in fig1‐A, page 1252 (Ahmed 2000). The mean BP change during first 24 hours (single point @ day 1;combining the 2 doses) was: Nimodipine (n=173): SBP ‐13.62±21.8; DBP ‐9.7±11.62 Placebo (n=92 ): SBP ‐4.8±25.7 ; DBP ‐3.6 ±13.7 SD of the change was not reported, it was imputed from end point Heart rate data: NR
Notes	Funding: Bayer AG Dates of conducting the trial: April 1989‐January 1990
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	use of a predetermined randomization list
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

ISIS‐1 1986.

Methods	Multi‐centre (245‐International ) Open‐label Method of randomization: based on a computer ?generated randomization list Concealment of allocation: by a 24 h direct line telephone service Duration of treatment: 7 days Follow‐up: 12 Months
Participants	16,027 patients within 12 hours after onset of symptoms of suspected AMI   Inclusion criteria: Patients with suspected myocardial infarction* within 12 hour of the onset of symptoms, not already on beta‐blockers or verapamil and with no clear indication for or contraindication to beta‐blockade (heart rate persistently below 100 mm 50 bpm, SBP <100 mm Hg), severe heart failure of bronchospasm)   * AMI was coded according to ECG findings as “probable” when a total elevation of ? 6 mm in leads V1‐V3 or V4‐V6 or of ? 2 mm in leads I and  AVL or ? 3 mm in leads II, III and aVF; or “possible” when less extreme ST elevation but with some other abnormality such as ST depression T‐inversion, pathological q‐waves; or normal.   Attrition data: Screened: Total randomized patients: 16,027, Atenolol:n= 8037, Control: n= 7990 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: 69 atenolol, 60 control The completeness of follow‐up was 99.3 in hospital and 97.7% to Jan 1 1985. Baseline characteristics: Age (years): Atenolol (A): 58.8± 8.96, Control (C): 58.9± 8.93 Females: A: 23%, C: 23% Mean delay (hours): A: 5 ± 2.69, C:  5 ±2.68 Mean SBP (mm Hg): A: 145 ± 26.89, C:  144.9 ±26.82 Mean heart rate/min: A: 79.2 ± 17.93, C:  144.9 ±17.87 Medical history Atenolol: Previous MI 1295 (16%), diabetes 6% Control: Previous MI 1327(17%), diabetes 6%
Interventions	Atenolol vs. no treatment (control) Drug regimen: Atenolol: 5 mg IV injection over 5 minutes and a second dose after 10 minutes (if no contraindications developed). Then, 50 mg orally followed by other 50 mg 12 h later. Then, 100 mg/ day for 6 days or until discharge if earlier. Control: beta‐blockers were avoided in hospital unless clearly indicated (chest pain unresponsive to nitrates, control of hypertension) Other interventions: Physicians were free to use other treatments felt appropriate. The percentage of drug use during hospital is as followed: Atenolol group: IV atenolol  94%, some oral beta‐blocker 92%, diuretic 38%, iv nitrates 7%, calcium antagonists 9%, digitalis 11%, anti‐arrhythmics15%, inotropic agents 5% Control group: IV beta‐blocker 2%, some oral beta‐blocker 7%, diuretic 37%, iv nitrates 8%, calcium antagonists 17%, digitalis 14%, anti‐arrhythmics17%, inotropic agents 3%
Outcomes	Mortality:  obtained from figure 1, page 59, and table III page 60 (all‐cause mortality) Atenolol (n=8037): at 2day:121(0.015); at 10 day (7d):317 (0.039); eof (1 year): 866 (0.107)  Control (n=7990): 2day: 171(0.021); at 10 day (7d): 367(0.046); eof (1 year): 951(0.12) Total non‐fatal SAE: NR Blood Pressure change during first 24 h: NR Note: In this trial there is information about mortality according to baseline BP and age Heart rate change during first 24 h: HR
Notes	Funding: ICI pharmaceuticals Ltd Dates of conducting the trial: mid 1981‐ Jan 1985
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	computer‐generated randomization list
Allocation concealment?	Low risk	patients were randomized by a 24‐hour direct line telephone service
Blinding? All outcomes	High risk	open control
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients.
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients. The completeness of follow‐up was 99.3 in hospital
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients. The completeness of end of follow‐up was 97.7
Free of selective reporting?	Low risk	Mortality data reported according to methods

ISIS‐4 1995.

Methods	Multi‐centre (1,086; multinational). Double‐blind. Method of randomization: computer generated list Concealment of allocation: by central phone; baseline details were recorded onto computer generated list before a specific numbered trial treatment pack was to be allocated. The computer used a “minimisation” algorithm which limited chance differences between treatment groups in these baseline features. Duration of treatment: 28 days Follow‐up: 1 year
Participants	58,050 patients with definite or suspected AMI within 24 hours of onset   Inclusion criteria: Patients with definite or suspected AMI within 24 hours with no clear indications or contraindications for the study drugs. Attrition data: Total randomized patients: 58,050 Captopril : n= 29,028 Placebo captopril: n= 29,022 Isosorbide mononitrate : n= 29,018 Placebo nitrate: n= 29,032 Total withdrawals (discontinuation of drug): NR Total withdrawals due to adverse events: NR Total lost of follow‐up: By discharge: the report on mortality was for 99% of the total randomized population At day 35 : Captopril : 1.8%, Placebo : 1.6%, Isosorbide mononitrate: 1.7% Placebo : 1.8%   Baseline characteristics: There were no differences in across the randomized groups. Overall of the randomized patients, at entry: 79% had ST elevation, 40 % were within 6 hours, 28% were 70 or older, 74% were male, 2% had SBP < 100 mm Hg, 17% had previous MI, 92% was confirmed to have an AMI
Interventions	Captopril vs Placebo vs. Isosorbide mononitrate   Drug regimens: Captopril: oral captopril; initially, 6.25 mg, then after two hours 12.5 mg, after twelve hours 25 mg. Maintenance: 50 mg bid for 28 days. Nitrate: control‐release isosorbide mononitrate; Initially, 30 mg, then after twelve hours 30 mg. Maintenance: 60 mg for 28 days Placebo: no further details. Other interventions: Captopril group: IV nitrates 13,652 (47%), any nitrate 15,878 (55%), anti‐platelets 26,905 (93%), fibrinolytic 19,917 (67%), IV BB 2,578 (9%) Placebo captopril group: IV nitrates 13,662 (47%), any nitrate 15,880 (55%), anti‐platelets 26,941 (93%), fibrinolytic 19,783 (68%), IV BB 2,541 (9%) Isosorbide group: IV nitrates 13,652 (47%), any nitrate 15,878(55%), anti‐platelets 26,921(93%), fibrinolytic 19751(68%), IV BB 2,541(9%) Placebo isosorbide group: IV nitrates 13,662 (47%), any nitrate 15,880(55%), anti‐platelets 26,925(93%), fibrinolytic 19,949 (68%), IV BB 2,578 (9%)
Outcomes	Mortality:  (day 2* was obtained from text 672‐673, day 10: from page 672 figure 1 by blowing‐up graph and calculate cumulative mortality at day 10; day 35: obtained from text on page 672‐673) * The authors of this trial reported mortality as "0‐1 day mortality". Day 0= patients dead the same day they entered the trial and day 1= deaths on the day after randomization. Captopril: N=29,028                 2d: 549 (1.89%); 10d: 1536 (5.29%); end of treatment (35d‐N/A): 2088 (7.19%)  Placebo Captopril: N=29,022,   2d:593 (2.04%); 10d: 1644 (5.66%); end of treatment (35d‐N/A): 2231 (7.69%) Isosorbide: N=29,018      2d: 514(1.77); 10d: 1548 (5.33) graph; end of treatment (35d‐N/A): 2129 (7.34%) Placebo Isosorbide: N=29,032,   2d:628 (2.16); 10d: 1667 (5.74) graph; end of treatment (35d‐N/A): 2190 (7.54%) Total Non‐fatal SAE: NR Blood Pressure change during first 24 h: NR Heart rate change during first 24 h: NR
Notes	Funding: The study was funded by Bristol‐Myers Squibb and by Astra‐Hassle. Dates of conducting the trial : July 1991‐August 1993
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Method of randomization: computer generated list
Allocation concealment?	Low risk	Concealment of allocation: by central phone; baseline details were recorded onto computer generated list before a specific numbered trial treatment pack was to be allocated. The computer used a “minimisation” algorithm which limited chance differences between treatment groups in these baseline features.
Blinding? All outcomes	Low risk	Comment: neither in the original publication (Lancet 1995;345:669‐685) nor in the published protocol Am J Cardiol 1991;68:87D‐100D) it is clearly stated that this trial was actually a double‐blind trial. Based on the description of its design it is assumed that there was double‐blinding for the comparison captopril vs. placebo and isosorbide vs. placebo ( in the consent form ‐page 91D‐ there is a paragraph that says "These study tablets would be either active or inactive ones ‐neither we nor you would know which, because the information is kept on a secret code list at another hospital.." ) but not for magnesium vs. open control. However, it is stated that the administration of other non‐study nitrates was permitted openly as long it was for the first few days.
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	By discharge: the report on mortality was for 99% of the total randomized population
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Jaffe 1983.

Methods	Single‐site study (US) Open‐label Method of randomization: NR Concealment of allocation: NR Duration of treatment: 24 hours Follow‐up: until hospital discharge
Participants	114 patients with myocardial ischemia or infarction within 12 hours of randomization and SBP of 100 mm Hg or greater and HR less 120 bpm Attrition data: Total randomized patients: Nitroglycerin (N): n=57 Placebo (P): n=57 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost of follow‐up: 0 Baseline characteristics: Values expressed as mean ± SD for continuous data Age (years): N:61±2 P:60±2 Mean SBP: N:130±39 C:134.8±33.7 Mean DBP: N:82.4±19.44 P:89±28 Male/female: N:28/15 P:30/12 Time to randomization (hours): N:6±0.4 P:6.4±0.5 Previous AMI: (%) N: 19 P: 21
Interventions	Nitroglycerin (N): n=57 Placebo (P): n=57 Drug regimens: N: Intravenous infusion of nitroglycerin at initial dose of 10 mcg /min (adjusted to achieve targets*). P: placebo (5 % dextrose and water) no further details *Target: either 10% reduction in SBP or achieve less than 95 mmHg at maximum dose of 200 mcg/min. 80% of patients achieved the target and it was achieve at mean of 1.74 ± 0.6 hours after starting the infusion. The mean dose of nitroglycerin to obtain target was 57± 21 mcg/ min. Mean dose administrated: NR Co‐interventions: Other than study treatment both groups were treated identically.
Outcomes	The results of this trial were reported in two papers (Jaffe 1983 and Roberts 1983). Mortality is not reported in either publication. However, in a review by Yusuf 1988 it is reported (based on a personal communication between Yusuf and Jaffe) that the mortality rate was Nitroglycerin =4/57, and Placebo=2/57. We could not confirm this information. Non‐fatal SAE: NR Individual SAE: NR Blood Pressure: Data was obtained from graphs in fig 1, page 455. The calculated weighted mean BP change during first 24 hours was: Nitroglycerine (n=43): SBP‐9.63±NR; DBP‐8.67±NR Placebo (n= 42): SBP‐8.9±NR ; DBP‐8.7±NR Standard deviation of change was not reported Heart rate: NR Heart rate: NR
Notes	Funding: National Institutes of Health
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"Patients were randomised", no further details
Allocation concealment?	Unclear risk	Not reported
Blinding? All outcomes	High risk	Not blinded
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Unclear risk	The results of this trial were reported in two papers (Jaffe 1983, Br Heart J 1983;49:452‐60 and Roberts 1983, The American Journal of Medicine 1983 June 27: 45‐52). Mortality is not reported in either publication. However, in a review by Yusuf 1988 it is reported (based on a personal communication between Yusuf  and Jaffe) that the mortality rate was Nitroglycerin =4/57, and Placebo=2/57.
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Unclear risk	potentially free of selective reporting as it was funded by NIH

Johannessen 1987.

Methods	Single‐site study (Norway) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 10 days Follow‐up: 10 days
Participants	40 patients within 6 hours after onset of symptoms of AMI   Inclusion criteria: Any age if admitted to the hospital within 6 hours after onset of symptoms of a suspected first myocardial infarction* and with and ECG indicating anterior location. *Diagnosis of AMI was based on a standard 12 leas EGG with new ST segment elevation > 1 mm in lead I and aVL or more than 2 mm in precordial leads.   Exclusion criteria: Heart rate below 50 bpm, SBP < 100 mm Hg, clinical signs of left ventricular failure, bronchial obstruction  or any degree of heart block; patients on digitalis B‐blockers, calcium antagonists or other cardioactive drugs   Attrition data: Screened: NR, Total randomized patients: 40, Timolol: n=   20, Placebo : n=20 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR   Baseline characteristics: Except for heart rate and other specialized cardiac indexes, the baseline characteristics were not reported Heart rate: Timolol: 77 ± 15, Placebo: 82  ± 14
Interventions	Timolol vs. Placebo   Drug regimen: Timolol: 1 mg IV bolus, repeated after 10 minutes. Then, an infusion of 0.6 mg / hour for 24 hours; followed by oral timolol 10 mg bid for 10 days  Placebo: Isotonic saline as placebo and placebo tablets. No further details Other interventions: Concomitant treatment which was thought to influence contractility was avoided. Furosemide, oxygen, morphine, and propoxyphene were allowed. All patients ventricular thrombus were treated with oral warfarin. Mean dose of furosemide during the first 4 days (mg/day): Tmolol group: 58±30 Placebo group: 50.5±60
Outcomes	Mortality:  obtained from text, page 153 Timolol: 2day:1/20 ; 10day:2/20   Placebo: 2day:0/20 ; 10day:0/20 Total non‐fatal SAE: NR Blood Pressure change within first 24 hours: NR Heart rate change within first 24 hours: NR
Notes	Funding: NR Dates of conducting the trial : NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomly assigned", no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods

Jugdutt 1983.

Methods	Single‐site study (Canada) Open‐label Method of randomization: NR Concealment of allocation: NR Duration of treatment: 24 Follow‐up: 12 days
Participants	22 patients with first anterior, transmural AMI within 6 hr of onset Inclusion criteria: Patients needed to have adequate 2D‐echo examination. Exclusion criteria: Patients with heart block or cardiogenic shock or history of past infarction or heart failure. Baseline characteristics for the two randomized groups: Nitroglycerin (N ): n=11 Control (C): n=11 Continuous values are expressed as mean ± SE Age(years) N: 59±3 C:54±4 Onset of pain to admission ( hours) N: 3±0.3 C: 2.7±0.7 Onset of pain to infusion ( hours) N: 5.9±0.4 C: 5.4±0.3 History of hypertension (n) N: 2 C: 1 SBP(mmHg) N: 130±6 C: 129±6 DBP(mmHg) N: 82±3 C: 83±4
Interventions	Nitroglycerin vs. Control Drug regimen: N: nitroglycerin IV infusion at initial rate of 5 mcg /min and increased by 5 to 20 mcg/min every 5 min in the first 30 min until MAP was reduced by 10 % ot its control value but not less than 80 mm Hg. The infusion was maintained for at least 24 hours. C: control patients receive 5% dextrose at constant rate of 1 ml/min for 24 hours. The average dose to lower MAP by 10 % was 29± 6 mcg/min and was achieve within 30 min. The average duration of infusion was 38.8 ± 4 hours Other interventions: All patients received nasal oxygen and intravenous morphine and were continued on lidocaine infusion ( 1 mg/min) throughout the study treatment. N: 2 patients received furosemide and digoxin C: 7 patients received furosemide and 2 digoxin.
Outcomes	Mortality: obtained from table 1, page 1267 Nitroglycerin, n= 11: 2day: 0; 10day: 1; eof( ): N/A Control, n=11: 2day: 0; 10day: 2; eof( ): N/A Non‐fatal SAE: obtained from page 1269 Nitroglycerin, n= 11: 0 Control, n=11: 2 (1 acute ventricular septal defect, 1 left ventricular failure) Blood Pressure; Data was reported as MAP over time. It was not possible to extract SBP or DBP data Heart rate: Data was obtained from graph in fig 2 page 1268. The calculated weighted mean HR change was: Nitroglycerin (n=11): 1±13 Placebo (n=11): 3±24 Standard deviation of change was not reported but imputed from end point
Notes	Funding: Canadian Heart Foundation and Special Services of the University of Alberta Hospital
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"Patients were randomized" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	High risk	no blinding was addressed or stated
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods

Jugdutt 1988.

Methods	Single‐site study (Canada) Single‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 48 hours Follow‐up: 43 months
Participants	310 patients with acute myocardial infarction within 12 hours of onset. IC: Patients with AMI (ECG 0.2 mV in two adjacent precordial leads, enzymes, clinical pain), SBP >100 mm Hg, HR < 120 bpm and within 12 hours of onset. EC: >75 years, HR < 55, bpm, killip 4, > 200/120, right ventricular infarction syndrome. Baseline characteristics for the two randomized groups: Nitroglycerin (N): n= 154 Placebo (P): n= 156 Continuous values are expressed as mean ± SD Age:(years) N:59±14 P:62±13 Male/female N:120/34 P:115/41 SBP:(mmHg) N:135±22 P:130±22 DBP:(mmHg) N:88±18 P:84±18 MAP:(mmHg) N:104±18 P:100±18
Interventions	Nitroglycerin vs Placebo Drug regimen: N: Intravenous infusion of nitroglycerin at initial dose of 5 mcg /min (upward titration up to reach target*) P: Placebo (IV 5% dextrose) no further details *The Targets were to reduce either: 10‐30% LVFP reduction ( but not more than 10% in MBP reduction) 10% SBP reduction (but not below 90 mmHg or 50 bpm HR increase) 10% MBP reduction (but SBP not below 90 mmHg) 10% MBP reduction (but MBP not below 80 mmHg) Treatment started within 4 hours in 22% of patients, 6 hour (43%), 10 hours (73%). Treatment lasted 48 hours (mean 39 hours) after randomization. Mean dose administrated: NR The mean dose of nitroglycerin to achieve target was 45 mcg/min (range 4‐192 mcg/min) Co‐interventions: during the first 48 hours all patients reveived nasal oxygen, iv morphine, continuous lidocaine (1mg/min). Thrombolysis was not mentioned. Acute in‐hospital phase: Nitroglycerin group: anti‐arrrhytmics 82, digoxin 38, furosemide 53, ibuprophen 21, beta‐blockers 24, calcium channel blockers 11, anticoagulants 13, anti‐platelets 4. Placebo group: anti‐arrrhytmics 76, digoxin 35, furosemide 54, ibuprophen 23, beta‐blockers 42*, calcium channel blockers 26*, anticoagulants 9, anti‐platelets 3. * p< 0.005 significance of difference comparing placebo vs. nitroglycerin groups.
Outcomes	Mortality: (reported on text, page 915) Nitroglycerin, n=154: 2d:NR ; 10d:7 ; eof(43 m):36 Placebo n=156: 2d:NR ; 10d:29 ; eof(43 m):50 Note: There is a discrepancy, in terms of number of deaths and number of patients randomized, between different publications of this trial. For example, in‐hospital deaths for nitroglycerin and placebo were reported as 14/154 (9%) and 29/124 (23%) respectively, in the abstract; but 22/154 (14%) and 41/156 (26 %), respectively, in the full‐paper publication. Non‐fatal SAE: NR Blood Pressure; Blood pressure data was reported as MAP (fig 1, page 910). There is no extractable data for SBP or DBP. The calculated weighted mean MAP change was: Nitroglycerin (n=154): ‐15.09±12.87 Placebo (n=156): ‐8.89±12.87 Standard deviation of change was not reported but imputed from end point Heart rate Data was obtained from graphs fig 1 page 910. The calculated weighted mean HR change was: Nitroglycerin (n=154):1.05±14.31 Placebo (n=156): ‐1.3±15 Standard deviation of change was not reported but imputed from end point
Notes	Funding: Canadian Heart Foundation, Ottawa, Ontario, Canada; By Alberta Heritage Foundation for Medical Research, Edmonton; and by Special Services of the University of Alberta Hospital.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"Patients were randomized" no further details
Allocation concealment?	Unclear risk	Not reported
Blinding? All outcomes	High risk	single‐blind. Comment: only double‐blind trials were considered yes to blinding
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Unclear risk	There was a discrepancy in the number of deaths and number of patients randomized (denominator) between two publications. (see selective reporting)
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	There was a discrepancy in the number of deaths and number of patients randomized (denominator) between two publications. (see selective reporting)
Free of selective reporting?	Unclear risk	There were two publications for this study: 1st (JACC 1985;5:447) and 2nd (Circulation 1988;78:906‐919). There was a discrepancy in the number of deaths and number of patients randomized (denominator) between these two publications. Whereas the denominator was the same for nitroglycerin group,154 patiens, in both publications; the denominator for the placebo group was very different in the two publications, 124 and 156 patients, respectively. Similarly, the number of deaths were different for nitroglycerin group; for example: mortality in‐hospital 14 and 22 respectively

Limburg 1990.

Methods	Single‐site study (Netherlands) Double‐blind Method of randomization: randomization tables by manufacturer Concealment of allocation: NR Duration of treatment: 14 Follow‐up: 6 months
Participants	26 patients within 24 hours after onset of symptoms of acute stroke Inclusion criteria: Supratentorial ischemic stroke with hemiparesis Exclusion criteria: Lacunar syndromes , serious underlying diseases, previous disabling strokes, using CCB, not being able to start within 24 hours Attrition data: screened /excluded at: NR Total randomized patients: Flunarizine : n=  12, Placebo : n= 14 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: Median Age, years (range): Flunarizine (F): 66  (56‐84), Placebo (P): 67.5 (22‐89) Sex m/f : F: 3/9, P: 6/8 SBP /DBP (mm Hg) : NR
Interventions	Flunarizine vs. Placebo Drug regimen: Flunarizine: as an IV bolus of  0.1mg/kg in 5% glucose solution, followed after 3 h by a continuous infusion of 0.3 mg/kg/24 h during 72 hours. Subsequently the drug was administered orally during 11 days as 10 mg/ day Placebo: inert vehicle administered similarly Other interventions: NR
Outcomes	Mortality:  obtained from text, page121, and table 2, page 122: total =8 Flunarizine: 2day: NR; 10day (7d):2/12; eof (6 m‐N/A):  3/12 Placebo: 2day: NR; 10day (7d):4/14; eof (6 m‐N/A):  5/14 Total non‐fatal SAE: NR Individual SAE: NR Blood Pressure data; NR Heart rate data : NR
Notes	Funding: Janssen Pharmaceutica, Tilburg, The Netherlands
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	randomization tables by manufacturer
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Lis 1984.

Methods	Multi‐centre (2 in Uk). Double blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 48 hours Follow‐up: 4 months
Participants	140 patients with clinical and electrocardiographic evidence of myocardial infarction occurring within 24 hours of admission. Exclusion Criteria Isolated peripheral hypoperfusion, outside 30‐75 years range, SBP?95 mm Hg, HR< 50 or >130 bpm, heart disease other than CAD, already receiving vasodilators, and symptoms > 12 hour prior to admission. Total withdrawals: (discontinuation of drug) Nitroglycerin (n=64): 17 ‐ 12 due to hypotension, 1 ventricular fibrillation, 1 due to SVT, 3 phlebitis. Placebo (n=76): 8 ‐ 4 due to hypotension, 1 cardiac arrest, 1 heart block, 2 phlebitis. Total lost of follow‐up: attendance at the 4‐month follow‐up was 77% . Baseline characteristics for the two randomized groups: Nitroglycerin (N ): n= 64 Placebo (P): n= 76 Age (years): NR Mean SBP (mmHg): N: 129 P: 136 Mean DBP (mmHg): N: 83.8 P: 85.5
Interventions	Nitroglycerin (N): n= 64 Placebo (P): n= 76 Drug regimen: N: Intravenous ntiroglycerin infusion of 10 mcg/ min (adjusted according to hemodynamic target*). P: placebo (10 % ethanol solution), no further details * bp target: if initial sbp was > 135= the aim was reduction of 30 mm Hg if initial sbp was 124‐134 = the aim was reduction of 15 mm Hg if initial sbp was 95‐119= the aim was reduction of 5 mm Hg Mean nitroglycerin dose administrated during the first 24 h was 70 mcg/min
Outcomes	Mortality: (reported on text, and table 1, page 181) Nitroglycerin, n= 64: 2day: NR ; 10day: 3 ; eof(4m):5 Placebo, n= 76: 2day: NR; 10day:6 ; eof(4m):10 Note: The total deaths at day 10 reported in the original publication by Lis 1984 was 9. But, the distribution of these 9 deaths according to the allocation group was not reported. However, in Yusuf 1988 meta‐analysis the distribution at day 10 was reported as 3 for nitroglycerin and 6 for placebo. Giving the benefit of doubt we decided to keep those numbers as reported in Yusuf 1988. We also decided to keep that distribution since it was proportional to that reported in the original publication for the end of follow‐up. Thus, it should not affect the overall effect size Non‐fatal SAE: NR Individual SAE: NR (page ) Ventricular fibrillation: Nitroglycerin, n= 64: 1 Placebo, n= 76: 0 Blood Pressure; Data was obtained from graphs in figure 1 page 181. The calculated weighted mean BP change was: Nitroglycerin (n= 64): SBP‐17.64±13.2; DBP‐9.68 ±9.99 Placebo (n=76): SBP‐11.53±17.76; DBP‐6.51 ±12.21 Standard deviation of change was not reported but imputed from end point Heart rate Data was obtained from graphs in figure 1 page 181. The calculated weighted mean HR change was: Nitroglycerin (n= 64): 6.30±13.3 Placebo (n=76): 1.3 ±16 Standard deviation of change was not reported but imputed from end point
Notes	Funding: Not Reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"Of the original 250 patients, a total of 140 were finally selected for entry and randomised.." Np further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	double‐blind placebo controlled trial.
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Unclear risk	The total deaths at day 10 reported in the original publication by Lis 1984 was 9. But, the distribution of these 9 deaths according to the allocation group was not reported. However, in Yusuf 1988 meta‐analysis the distribution at day 10 was reported as 3 for nitroglycerin and 6 for placebo(page.1091).
Incomplete outcome data addressed? 30‐day mortality	Low risk	total lost of follow‐up was 23%
Free of selective reporting?	Unclear risk	The total deaths at day 10 reported in the original publication by Lis 1984 was 9. But, the distribution of these 9 deaths according to the allocation group was not reported

Marangelli 2000.

Methods	Multi‐centre (10 in Italy) Double‐blind Method of randomization: computer software Concealment of allocation: central coordinating center Duration of treatment:  24 hours Follow‐up:  90 days
Participants	90 patients within 4 hours after onset of symptoms of their first acute anterior myocardial infarction who already had given a thrombolytic treatment. Inclusion criteria: AMI defined as typical chest pain lasting > 30 min unresponsive to sublingual trinitroblycerin,  with  electrocardiographic signs of ongoing myocardial infarction (ST segment elevation > 0.1 mV in leads D1 and aVL and or > 0.2 mV in two or more precordial leads from Vi‐V6). Killip class I, with technically excellent two‐dimensional digital echocardiograms ( defined as a stop‐frame identifying > 85%  of the endocardial border). Exclusion criteria: Age > 75 years, pre‐existing  heart failure, previous MI, coronary surgery or angioplasty in the 6 months preceding the infarction, significant valvular disease, cardiomyopathy, congenital heart disease, contraindications to thrombolytics or heparin, killip class > 1 signs of hypoperfusion; heart rate < 60 b/min, SBP < 100  mm Hg, arrhythmias, refuse to give consent, chronic obstructive  pulmonary disease, obesity, pacemaker. Attrition data:  Total randomized patients: 90, Verapamil (V ): n=44, Placebo (P): n=44 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost of follow‐up: 2 (not available for analysis, no further details) Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years) V: 55.8 ±9.7, P: 57.8±10.6 SBP (mm Hg): NR Time to treatment (min): V:169 ±65, P: 141±83 Time to thrombolysis (min): V:178±68, P: 153±80 Medical history: Verapamil: hypertension 15, diabetes 12, smoke 24, hypercholesterolemia 8,  angina 26 Placebo: hypertension 16, diabetes 8, smoke 22, hypercholesterolemia 12,  angina 28
Interventions	Verapamil vs. Placebo Drug regimen: Verapamil: 5 mg iv bolus  follow by an infusion of 2 mcg/kg/min over 24 hours. This treatment was started before beginning thrombolytic therapy Placebo : identical ( in appearance) to verapamil  formulation Other interventions: All patients received thrombolysis: tissue‐type plasminogen activator‐rt‐PA (15 mg IV bolus followed by an infusion of 0.75 mg/kg up to a maximum of 50 mg over 30 min, and a further infusion of 0.50 mg/kg up to a maximum of 35 mg over 60 min); Heparin 500 IU IV bolus, followed by an infusion of 1000 IU/hour (1200 IU/hour in >80 kg patients);and aspirin 160 ‐325 mg / day. Verapamil group: ACEi 26, beta‐blockers 9, CCB 3, nitrates 20 , diuretics 13, cortisone 1. Placebo group: ACEi 27, beta‐blockers 15, CCB 7, nitrates 26 , diuretics 19, cortisone 0
Outcomes	Mortality:  obtained it from text , page 340 Verapamil : 2day:NR ; 10day (Hospital):1/44 ; eof (90 days):NR  Placebo: 2day:NR ; 10day (Hospital):0/44 ; eof (90 days):NR Total non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR Heart rate change during first 24 hours: NR
Notes	Funding: NR Dates of conducting the trial: March 1995‐May 1995
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	computer software
Allocation concealment?	Low risk	central a coordinating centre
Blinding? All outcomes	Low risk	double‐blind
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods

MIAMI 1985.

Methods	Multi‐centre (104‐International ) Double‐blind Method of randomization: computer generated number Concealment of allocation: number coded sealed envelopes Duration of treatment: 15 Follow‐up: 15
Participants	5778 patients within 24 hours after onset of symptoms of suspected AMI   Inclusion criteria: Men and women 75 years old or less with chest pain of acute onset and suspicion of acute myocardial infarction of at least 15 min of duration, ECG signs and symptoms indicating AMI* within the last 24 hours. * AMI was diagnosed if 2/3 criteria fulfilled: 1) chest pain with a duration of at least 15 minues; 2) two values above the normal range for enzymes and; 3) appearance of new Q waves or loss of R waves or ST‐segment elevation followed  by T‐wave inversion in at least 2 electrodes in a 12‐lead standard ECG.   Exclusion criteria: current treatment with beta blockers or CCB within 48 hours , HR less than 65 bpm, SBP less than 105 mm Hg , left ventricular  heart failure, signs of poor peripheral circulation, 2 or 3 degree AV block, other serious diseases influencing short‐term prognosis.   Attrition data: Screened: NR, total randomized patients: 5778, Metoprolol : n= 2877, Placebo : n= 2901 Total withdrawals (discontinuation of drug): Metoprolol: 441 (15.3%), Placebo : 401 (13.8%) Withdrawals due to adverse events: Metoprolol: 373 (13 %), Placebo : 214 (7.4%) Total lost to follow‐up: 2, Metoprolol : 0, Placebo : 2 Baseline characteristics: Note: continuous variables are expressed as median Age (years): Metoprolol  (M): 60, Placebo (P): 60 Gender male/female (%): Metoprolol  (M): 78/ 22, Placebo (P):77.2/ 22.8 Time to treatment: (hours): M: 6.8, P:  6.5 Heart rate: M: 80, P:  80 SBP (mm Hg): M: 140, P: 140 No. of patients with AMI; other diagnosis (%) Metoprolol (M): 2041 (71%); 836 (29%) Placebo (P):2017 (69.5%); 883 (30.5%) Patients with AMI: anterior; inferior; other (%) Metoprolol (M): 974 (48%); 786 (38 %); 281(14%) Placebo (P): 958 (47%); 800 (40 %); 259 (13%) Medical history (%) Metoprolol group: previous infarction 459 (16.0), angina pectoris (27.7), congestive heart failure (3.9), hypertension (13.6), diabetes (6.7), smokers (50.6) Placebo group: previous infarction 467 (16.1), angina pectoris (29.5), congestive heart failure (3.6), hypertension (14.1), diabetes (7.6), smokers (52.1)
Interventions	Metoprolol  vs. Placebo Drug regimen: Metoprolol : 5 mg x 3 IV bolus with a 2‐minute intervals (HR  should be > 49 bmp,  and SBP > 99 mm Hg and dyspnea or cold sweating not worsening ); then 100 mg orally  every 6 hours for 2 days,  and then 200 mg every 12 hours for 13   or 14 additional days. Placebo: no further details Other interventions: The general management of patients was according to local practice. Other B‐blockers were not allowed. Thrombolysis was not reported Metoprolol group: antiarrhythmic 552(19.2%), cardiac glycosides391 (13.6%)*, diuretics 1233 (42.9%), narcotic analgesics 1269(44.1%)*, nitrates 1527(53.1%), CCB 252 (8.8%)*, atropine 205(7.1%)*, simpathomimetics 163 (5.7%) Placebo group: antiarrhythmic 624 (21.5 %), cardiac glycosides 482 (16.6%)*, diuretics 1206 (41.6%), narcotic analgesics 1431(49.4%)*, nitrates 55.3(1602%), CCB 343 (11.8%)*, atropine 135 (4.7%)*, simpathomimetics 121 (4.2%) * p<0.001
Outcomes	Mortality:  obtained from figure 1, page 204 Metoprolol (n=2877): 2day:29 (0.01) ; 10day:100 (0.035)  ; eof( 15 days):  123 (0.043) Placebo (n=2901): 2day:41 (0.014) ; 10day:110 (0.038)  ; eof(15 days ): 142 (0.049) Total non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR Heart rate change during first 24 hours: NR
Notes	Funding: the funding source is not reported but trial medication was supplied by Astra Co, Dates of conducting the trial : November 1982‐ March 1984
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	computer generated system
Allocation concealment?	Low risk	number coded seal envelopes
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

MILIS 1984.

Methods	Multi‐centre (5‐USA ) Single‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 10 days Follow‐up: 36 months
Participants	269 patients within 18 hours after onset of symptoms of AMI   Inclusion criteria: Less than 76 years, typical chest pain for more than 30 min, ECG new  Q waves, > o.1 mv ST elevation, or depression or both, or LBBB, OR idioventricular  rhythm .   Exclusion criteria: Less than 18 years, pregnant, cardiogenic shock, serious illness, pacemaker, AMI within 2 weeks, receiving nitrates or BB that could not be discontinued for more than 72 hours. Contraindications to BB   Attrition data: Screened: 7597, Not eligible for this trial 6,718, Of those eligible (n=879, group A), 178 were excluded for physician refusal, 140 for patient refusal and 161 other reasons. (total excluded=479) Total randomized patients: 400 Propranolol: n= 134 Placebo : n= 135 Hyaluronidase: n=131* * this group will not be considered any further Baseline characteristics: Mean Age (years): Propranolol  (P): 54.9, Placebo (Pbo): 54.6 Male (%): P: 72.4, Pbo: 74.1 Location of ischemia/infarction: anterior, inferior, other Propranolol  (P): 54.5%, 40.9%, 4.6% Placebo (Pbo): 61.7%, 35.3%, 3% ECG findings P: LBBB 0.8%, ST elevation 82.6%,  ST depression 13.6% Pbo: LBBB 1.5%, ST elevation 84.2%,  ST depression 12.8%   Mean time between onset and treatment (hr) P: 8.5,  Pbo: 9.3 Mean HR (bpm): P: 79.6, Pbo: 81.3 BP : NR Medical hx Propranolol  group: Smokers  (47%), higher education (62%), Regular drinkers (18.1%) HTN (50.8%), PREVIOUS AMI (14.9%), angina(42.5%), CHF (3%), diabetes (20.2%), previous cardiac arrest (1.5%), previous cardiac surgery (3.7%), arrhythmias (7.5%), Placebo : Smokers (53%), higher education (60.6%), Regular drinkers (24.2%) HTN (37%), PREVIOUS AMI (14.1%), angina(35.6%), CHF (6.7%), diabetes (18.5%), previous cardiac arrest (0.0%), previous cardiac surgery (5.9%), arrhythmias (7.4%)
Interventions	Propranolol  vs. Placebo Drug regimen: Propranolol: 0.1 mg/kg over 6 minutes, after 3 hours a second bolus of  0.025 mg /kg. Then, after 3 more hours oral propranolol was given at a dose of 20 mg to up to 600 mg /per day (to keep HR  between 45 and 60 bpm and SBP > 90 mm Hg). Treatment continued for 7 days, and then tapered to one half  during the 8th and 9th days and discontinued on the 10th day Placebo: No further details Other interventions: Medication within 3 weeks Propranolol group: Beta‐blockers (24.6%), nitrates(17.9%), sl. nitroglycerin (23.1%), anti‐arrythmics (3%), anticoagulants (0%), digoxin (9.7%), diuretics (29.9%), oral hypoglycemics (4.5%) Placebo group: Beta‐blockers (17.8%), nitrates(17%), sl. nitroglycerin (20%), anti‐arrythmics (1.5%), anticoagulants (3%), digoxin (7.4%), diuretics (18.5%), oral hypoglycemics (2.2%)
Outcomes	Mortality:   Propranolol group: 2day:NR   10day: 4/134 (obtained from table 1, page 40F ( Am J Cardio; 57) eof( at 36 months): 24/134 (obtained from text, (NEJM 1984; 311: p.221)   Placebo group: 2day:NR 10day: 8*/135 ( *The total, in‐hospital or 1‐month, deaths for both groups was 12 [NEJM 1984; 311: p.223]. If 4 deaths occurred in the propranolol group ( Am J Cardio 1986; 57:38F‐42F) It is assumed that the remaining 8 deaths occurred in the placebo group) eof( at 36 months): 20/135 (obtained from text, (NEJM 1984; 311: p.221)   Total non‐fatal SAE: NR Blood Pressure: NR during first 24 hours   Heart rate: NR during first 24 hours
Notes	Funding: Ayerst Laboratories Dates of conducting the trial : August 1 1978‐ February 1, 1983
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomized" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	High risk	"single‐blind" Comment: only double‐blind trials were qualified as "yes" for blinding
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Unclear risk	Mortality data up to day 10 was selectively given only for propranolol group (4 deaths). Mortality for placebo group was taken from the worse case scenario assuming 8 deaths in the placebo group happened within the first 10 days after randomization)
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Free of selective reporting?	High risk	Mortality data up to day 10 was selectively given only for propranolol group

Mitchell 2002.

Methods	Multi‐centre (21‐USA) Open‐label Method of randomization: computer‐driven response system and generation list. Randomization was stratified by type of ischemic syndrome. Concealment of allocation: telephone system ensured proper sequence allocation, providing security against potential randomization bias. Duration of treatment: up to 36 hours Follow‐up: 6 weeks
Participants	108 patients within 12 hours after onset of symptoms of AMI   Inclusion criteria: Acute transmural myocardial infarction, TMMI^, (ie, ST elevation) or unstable angina/non‐Q‐wave myocardial infarction (UA/NQW)^. Having at least one B‐blockade relative contraindication such as left ventricular dysfusnction, mild congestive heart failure, history of bronchospastic airway disease (without bronchospasm), prolonged ECG‐PR interval, controlled  diabetes mellitus, hypotension (SBP >100 mm Hg) bradicardia (>55 bpm), concomintant AV node‐blocking CCB‐ diltiazem verapamil). ^TMMI was defined as angina type chest pain of > 20 minutes duration, unresponsive to sublingual nitroglycerin and ST elevation of > 10 mm in at least 2 of the 3 inferior leads (II, III, avF) or in at least 2 contiguous precordial leads (V1‐V6) or in leads I, aVL. UA/NQW  was defined as chest pain at least 5 minutes, unrelieved by sl nitroglycerin and either positive CK‐MB  or concomitant  ST‐depression (>0.5 mm). T‐wave inversion of ? 1 mm or transient (< 20 minutes) ST segment elevation.   Exclusion criteria: Bradycardia (<55 bpm), hypotension (SBP < 100 mm Hg) , prolonged PR (>0.30 seconds), 2‐3 degree AV block, severe congestive heart failure, acute bronchospatic episode,pregnancy; atrial fibrillation, wolff‐parkinson‐white syndrome, permanente ventricular pace makers, planned surgical revascularization,  drug or alcohol abuse, serious advanced illness; receiving B‐blockers within 24 hours, CCB within 48 hours.   Attrition data: Screened: NR, total randomized patients: 108, Esmolol: n= 55, No esmolol or Control group: n= 53 Total withdrawals (discontinuation of drug): 12 (all in esmolol group) Withdrawals due to adverse events: 12 (all in esmolol group‐hypotension bronchospasm, AV block, bradycardia) Total lost to follow‐up: NR Baseline characteristics: Mean age in years (range): Esmolol (E): 58.1 (31.7‐88.3), No esmolol or Control group (X): 60.4 (35‐84.4). Male: E: 26%, X: 40% Race: white, black, other (%): E: 35, 18, 2 ; X: 35,11,7 Heart rate, bpm, (range) E: 81.0 (58‐125), X :78.7 (56‐109) SBP (mm Hg) : E: 128.1 (95‐200),  X: 130.7 (97‐197) ECG criteria: ST elevation, non‐ST elevation (%): E: 23, 32 ; X: 21, 32
Interventions	Esmolol vs. Control group (no esmolol group) Drug regimen: Esmolol: 500 mcg/ kg IV over 1 minute. Followed by infusion of 50 mcg/kg/min (titration in increments of 50 mcg/kg/min every 5 to 15 minutes) 10 mcg /min* for 16 to 30 hours^; 30 minutes before discontinuation oral metoprolol was started (12.5 to 100 mg for 6 weeks No esmolol or Control group: received the standard medical therapy? but no IV beta‐blocker or Oral beta‐blockers during the first 24 hours. Then, oral metoprolol (12.5 to 100 mg) was started and continued for minimum of 6 weeks. ^ until a peak dose of 300 mcg/kg/min or end safety point (developing of new rales, pulmonary edema, 2‐3 degree AV block, shock, bronchospastic episode. Other interventions: All patients received standard medical therapy (including thrombolytic therapy, intravenous heparin, aspirin, nitrates, narcotics), Oral verapamil or diltiazem were allowed only to the control group but other CCB were used in esmolol group (12 pts) and in the Control group (9 pts)
Outcomes	Mortality:  obtained from table II , page 110 Esmolol group (E) At 2day:NR ; 10day (in hospital):2/55 ; eof(6 weeks ):  2/55 No esmolol or Control group (X) At : 2day: ; 10day (in hospital):1/53 ; eof(6 weeks ):  1/53 Total non‐fatal SAE: NR Blood Pressure change in first 24: NR Heart rate change in first 24: NR
Notes	Funding: Not stated, but Baxter Pharmaceutical Products Inc, New Providence, NJ, provided the study medication Dates of conducting the trial: 1975
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	computer‐driven response system and generation list. Randomization was stratified by type of ischemic syndrome.
Allocation concealment?	Low risk	by central telephone
Blinding? All outcomes	High risk	blinding was not addressed or stated
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Free of selective reporting?	Low risk	Mortality data reported according to methods

Muller 1984.

Methods	Multi‐centre (6‐US ) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 14 Follow‐up: 6 months
Participants	181 patients within 6 hours after onset of a threatened AMI Inclusion criteria: Diagnosis was based on chest pain for more than 45 minutes and new or presumably new ST segment elevation or depression of at least 0.1 mV or new Q waves of at least 30 msec width and 0.2 mV depth in at least two of three diaphragmatic leads (II,III, aVF) or at least two of six precordial leads, or in I and aVL. Exclusion criteria: Left bundle branch block, younger than 21 or older than 80 years, SBP < 110 mmHg, previous major illnesses, CABG or AMI (within 21 days), childbearing potential, inability to cooperate, previous participation in trial. Attrition data: Total screened: 3143, total randomized patients: 181, Nifedipine (N ): n=93, Placebo (P): n= 88 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: Nifedipine: 15/93, Placebo : 13/88 Total lost of follow‐up: NR Baseline characteristics: Except for mortality, baseline and all results are given as per protocol, not as per ITT. According to per protocol there was Not significant difference between the two groups in any variable. For example, History of previous MI: Nifedipine: 16/89 (18%) Placebo: 14/82 (17%)
Interventions	Nifedipine vs. Placebo Drug regimen: Nifedipine: 2 capsules of 10mg (total 20 mg) every 4 hours for 14 days. Placebo: two identical‐appearing placebo capsules, no further details. Therapy was reduced to 10 mg or eliminated if severe adverse effects. For example, SBP < 85; or 30 mm Hg of more from baseline. Other interventions: In general nitrates and b‐blocking drugs would be avoided. For patients with threatened MI could receive them 24 or more hours after the qualifying episode of angina.
Outcomes	Mortality:  obtained it from table 4, page 744 Nifedipine: 2day: NR ; 10day (14 day):7/93  ; eof: not aplicable  Placebo: 2day: NR ; 10day (14 day):2/88  ; eof: not aplicable Total non‐fatal SAE: NR Blood Pressure data on first 24 hours: NR Heart rate data on first 24 hours: NR
Notes	Funding: Pfizer Pharmaceuticals Inc., New York. NY. Dates of conducting the trial : initiated in 1979
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were assigned randomly"; no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	High risk	Baseline characteristics not given for all randomized patients

Nabel 1991.

Methods	Single‐site study (US) Double‐blind Method of randomization: performed by the manufacturer by a four block‐size randomization code Concealment of allocation: All participants (patients, physicians nurses and pharmacist) had no information on the study medication. Code was broken after all data were collected Duration of treatment: 3 months Follow‐up: 3 months
Participants	38 patients within 6 hours after onset of symptoms of AMI   Exclusion criteria: Chest pain relieved by nitroglycerin, less than 20 min, age >75, contraindication to thrombolytic or captopril, SBP / DBP  > 180 / 110 mm Hg, current therapy with cytotoxic drugs or serious advance illness, pregnancy , prior Q wave infarction.   Attrition data: Screened: NR, Total randomized patients: 38, Captopril: n= 20, Placebo : n= 18 Total withdrawals (discontinuation of drug): 1 captopril Withdrawals due to adverse events: 1 from captopril group ( due to hypotension SBP < 85 mm Hg) Total lost to follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years) : Captopril (C): 52.9 ± 13, Placebo (P): 56.4± 10 Gender M/ F: C:19/1, P:12/6, Time to rt‐PA (hours): C: 3 ± 1.8, P: 3 ± 1.3 SBP/ DBP (mm Hg) NR
Interventions	Captopril vs. Placebo Drug regimen: initially, intravenous captopril , 2 mg, over 1 min. Then, additional 8 mg (if SBP > 85 mm Hg, after 2 minutes). Then, 3 to 4 hours after, oral treatment 12.5 mg twice and titrated up to 25 and 50 mg bid. If tolerated. Continuing medication for 3 months. Placebo: No further details   Other interventions: given according to group (%): Captopril group: beta‐blockers 5 (25), CCB 6 (30), diuretics 3 (15), anti‐arrhytmic 1 (5), nitrates 9 (45), Placebo group: beta‐blockers  8 (44), CCB 5 (28), diuretics 1(6), anti‐arrhytmic 0(0), nitrates 9 (50).
Outcomes	Mortality:  obtained from text, page 470 Captopril : 2day: NR ; 10day: 0/20  ; eof(3 months ):  N/A Placebo : 2day: NR ; 10day: 1/18  ; eof(3 months ):  N/A   Total non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR   Heart rate change during first 24 hours: NR
Notes	Funding: Bristol‐Myers Squibb Dates of conducting the trial : March 1988 and May 1989
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"randomization was performed by the manufacturer by a four block size"
Allocation concealment?	Low risk	using a randomization code not to be broken until all data were collected
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Natale 1999.

Methods	Single‐site study (Italy) Double‐blind Method of randomization: a list by sets of 10 (5 for verapamil and 5 for placebo), no further details Concealment of allocation: NR Duration of treatment: 6 months Follow‐up: 6 months
Participants	70 patients within 12 hours after onset of symptoms of the first anterior acute MI who did received IV thrombolytic treatment within 6 h of onset. Inclusion criteria: Aged 18‐75 years, good echocardiographic window. diagnosis of AMI was based on typical pain lasting ? 30 minutes and ST‐segment elevation ? 1 mm in at least two contiguous precordial leads and or in DI  and aVL. Exclusion criteria: Overt heart failure, ejection fraction < 45%, SBP < 90 mm Hg, second or third‐degree AV block or SA block, left ventricular hypertrophy, congenital or valvular heart disease, liver or kidney failure, treatment with BB or CCB on admission, inability to take part in the study Attrition data: Total randomized patients: 70, Verapamil: n= 35, Placebo: n=35 Total withdrawals (discontinuation of drug): 24, Verapamil: 14, Placebo: 10 Withdrawals due to adverse events: 11 (not reported according to randomization group) Total lost of follow‐up: 0 Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years): Verapamil (V): 60 ± 10.6, (26‐74, range); Placebo (P): 56±9.1, (39‐73, range) Males/females: V: 30/5, P: 30/5 Pre‐thrombolysis time (hours): V: 2.8 ±1.5, P: 2.5 ±1.1 SBP (mm Hg): V: 136±19, P:  137±24 DBP (mm Hg): V: 80±13, P:  83±14 HR (bpm): V: 77±13, P:  76±13 Medical history: Verapamil: Smokers 15, diabetes 4, hyperlipedaemia 9, hypertension 14, ischemic heart disease 9, angina 9, PTCA 1. Placebo: Smokers 25*, diabetes 3, hyperlipedaemia 6, hypertension 10, ischemic heart disease 14, angina 8, PTCA 0
Interventions	Verapamil  vs. Placebo Drug regimen: Verapamil: 5 mg / hour IV infusion for 24 hours, followed by oral administration of verapamil retard 120 mg  t.i.d. for 6 months Placebo: placebo, no further details Other interventions: All patients received IV accelerated recombinant tissue‐type plasminogen activator (100 mg). BB or other CCB was not permitted.
Outcomes	Mortality:  obtained from text, page 318 Verapamil: 2day: NR ; 10day: 2 (before discharged)  ; eof: not aplicable  Placebo: 2day:NR ; 10day: 0 (before discharged)  ; eof: not apllicable Total non‐fatal SAE: NR Blood Pressure; Data was obtained from graphs in Fig 1, page 320. The mean BP change during first 24 hours (up to hour 24) was: Verapamil (n=35): SBP= ‐15.40 ±14; DBP= ‐9.53 ± 9.35 Placebo (n=35): SBP=‐12.59 ± 17.14; DBP=‐9.05±12.47 Heart rate: not reported during up to the first 24 hours.
Notes	Funding: Not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	use of a randomization list
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Norris 1978.

Methods	Single‐site study (New Zealand) Open‐label Method of randomization: NR Concealment of allocation: by the envelope method Duration of treatment: 27 hours Follow‐up: hospital stay
Participants	43 patients within 4 hours after onset of symptoms of AMI   Inclusion criteria: Men and women aged up to 65 admitted within 4 hours of the onset of  AMI ( patients were subdivided into two subcategories: Group 1‐patients with ST‐segment depression 1‐3 mm in any lead or T‐wave inversion without ST abnormalities. Group 2‐ patients with normal ECG ) Exclusion criteria: Unstable angina (repeated attacks of pain over the last days or weeks); patients with contraindications to propranolol (history of cardiac failure, bronchial asthma, heart rate < 60 bpm), taking b‐blockers within 72 hours, CK elevation caused by electrical defibrillation or muscular injections; patients with ST‐segment elevation > 2 mm in precordial leads and AVF; pathological Q waves not due to a known previous infarct.   Attrition data: Screened: NR, total randomized patients: 43, Propranolol: n=  20, No propranolol (control) group : n= 23 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: Mean age, in years (range): Propranolol (P) :54 (45‐65), No propranolol (control) group (X) : 51(32‐65) Sex (M/F): P: 14/ 6, X: 18 / 5 History of previous infarction: P: 3 (15 %), X:  6(26 %) SBP/DBP (mm Hg) : Propranolol (P) : 144.54/ 90, No propranolol (control) group (X) : 144/88
Interventions	Propranolol   vs. No propranolol (control) group Drug regimen: Propranolol: 0.1 mg /kg over 10 min followed by a total of 320 mg orally over the next 27 hours   Other interventions: Oral furosemide or IV lignocaine were allowed.
Outcomes	Mortality:  obtained from text, page 908 Propranolol: 2day: 0/20; 10day (hospital): 0/20   No propranolol (control) group2day: 0/23; 10day(hospital): 0/23  Total non‐fatal SAE: NR Blood Pressure: Data was obtained from figure 1, page 908. The mean BP at end point (24 h) was: Propranolol (n=20): SBP ‐21.95 ±17.88 ; DBP‐11.69 ±14.22 Control (n=23): SBP ‐16.02 ±15.25 ; DBP ‐4.33 ± 15   Heart rate: Data was obtained from figure 1, page 908. The mean HR at end point (24 h) was: Propranolol (n=20 ):‐9.22 ±10.16 Control (n=23 ):2.21 ±12
Notes	Funding: Medical Research Council, National Heart Foundation of New Zealand
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"patients randomized" no further details, but probably well done as other publications of the same author clearly state methods
Allocation concealment?	Low risk	"by the envelope method"
Blinding? All outcomes	High risk	open control
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Data reported according to methods

Norris 1980.

Methods	Two‐site study (New Zealand) Open‐label Method of randomization: NR Concealment of allocation: by the envelope method Duration of treatment: 27 hours Follow‐up: hospital stay
Participants	62 patients within 4 hours after onset of symptoms of AMI   Inclusion criteria: Men and women aged up to 65 admitted within 4 hours of the onset of  AMI^. No history of bronchial asthma, SBP > 100 mmHg, heart rate greater than 60 /min and without breathlessness of basal rales.   ^Uncomplicated transmural infarction as evidence by a) typical history of prolonged chest pain with onset less than four hours previously and b) ST segment elevation in ECG greater than 2 mm in anterior chest leads, > 1 mm in II, III aVF or pathological Q waves.   Exclusion criteria: Patients who had had DC cardioversion (cardiac arrest), evidence of overt cardiac failure, hypotension, interstitial or pulmonary oedema Attrition data: Screened: NR, total randomized patients: 62, Propranolol: n=   33, control (No propranolol)  group : n= 29 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: Mean age, in years (range): Propranolol (P) :51 (31‐64), No propranolol (control) group (X) : 51(37‐65) Sex: M/ F: P: 31/ 2, X: 27/ 2 Position of the infarct : anterior, inferior, antero‐inferior,  subendocardial: P: 17, 13,0,3; X: 14,14,1,0, History of previous infarction: NR SBP/DBP (mm Hg): NR
Interventions	Propranolol vs. Control (No propranolol)  group Drug regimen: Propranolol: 0.1 mg /kg over 10 min followed by a total of 320 mg orally over the next 27 hours (40 mg at one,3, 7 11, 15,19,23, 27 hours after entry to the trial ( before each oral dose, SBP should be > 100 mm Hg and HR > 50 bpm) Other interventions:Oral furosemide or IV lignocaine were allowed.
Outcomes	Mortality:  obtained from table 2, page 619 Propranolol: 2day: NR; 10day (in hospital): 1/33 ; eof: NR  No propranolol (control) group : 2day: NR; 10day (in hospital): 0/29 ; eof: NR Total non‐fatal SAE: NR Blood Pressure change in first 24: NR Heart rate change in first 24: NR
Notes	Funding: Medical Research Council, National Heart Foundation of New Zealand Trial conducted between March 1977 and March 1979
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"patients randomized" no further details, but probably well done as other publications of the same author clearly stated method
Allocation concealment?	Low risk	"by envelope method"
Blinding? All outcomes	High risk	open label
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods

Norris 1984.

Methods	Multi‐centre study (4, New Zealand) Open‐label Method of randomization: NR Concealment of allocation: by the envelope method Duration of treatment: 27 hours Follow‐up: hospital stay
Participants	735 patients within 4 hours after onset of symptoms of AMI   Inclusion criteria: Patients under 70 years old  within 4 hours of the onset of  AMI (complaining of chest pain for more than 30 min) , provided no contraindications to beta‐blockers (bronchitis requiring bronchodilators, current treatment for cardiac failure, dyspnea or widespread chest rales, systolic blood pressure below 110 mm Hg or HR below 60 bpm. Exclusion criteria: Delayed ( more than 4 hours) and contraindications Attrition data: Screened: NR, Total randomized patients: 735, Propranolol: n=   364, No propranolol (control) group : n= 371 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: (±SD) Mean age, in years: Propranolol (P) :55 ±9, No propranolol (control) group (X):54±10 Sex (M/ F)P: 299 / 65, X: 289/ 82 History of previous infarction P: 94 /364 (25.8 %), X:  95 /371 (25.6 %) ECG findings on entry: pathological Q waves, ST‐ segment elevation, ST depression, T wave changes only, normal ECG P: 69,165.32,47,106 (29%), X:  77,147,47,56,115 (31%) SBP/DBP (mm Hg): Propranolol (P) : 144 ± 24 / 91± 14 , No propranolol (control) group (X) : 145 ± 25 / 91± 15
Interventions	Propranolol vs. No propranolol (control) group Drug regimen: Propranolol: 5‐8 mg over 5 min followed by a total of 320 mg orally over the next 27 hours (40 mg at one,3, 7 11, 15,19,23, 27 hours after entry to the trial (before each oral dose, SBP should be > 100 mm Hg and HR > 50 bpm). Note: mean dose given was 6.5 mg (IV), and 236 mg (oral). Control group: If oral beta‐blockers being taken at the time of entry to the trial they were usually continued during the acute phase of infarction. Other interventions according to study groups: Note: Oral furosemide or IV lignocaine were allowed. Propranolol:  Long‐term beta blocker continued 0, morphine or other analgesic 193, diuretic 61, pressor agent 10, digoxin 16, lignocaine 66, atropine 25, pacemaker 6. No propranolol (control) group : Long‐term b‐blocker continued 83, morphine or other analgesic 206, diuretic 50, pressor agent 8, digoxin 14, lignocaine 68, atropine 15, pacemaker 4
Outcomes	Mortality:  obtained from text, page 885 Propranolol: 2day: NR; 10day (in hospital): 15/364; eof : NR  No propranolol (control) group: 2day: NR; 10day (in hospital): 14/371; eof: NR Total non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR Heart rate change during first 24 hours: NR
Notes	Funding: Medical Research Council, National Heart Foundation of New Zealand Trial conducted during May 1981‐ March 1984
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"patients randomized" no further details, but probably well done as other publications of the same author clearly stated method
Allocation concealment?	Low risk	"by envelope method"
Blinding? All outcomes	High risk	blinding was not addressed or stated
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods

Owensby 1985.

Methods	Single‐site study (New Zealand) Open‐label Method of randomization: Stratified according to timing of entrance (< 4 and > 4 hours), no further details. Concealment of allocation: by sealed envelope method Duration of treatment: 2 days Follow‐up: During hospital
Participants	100 patients within 12 hours after onset of symptoms of suspected AMI   Inclusion criteria: Under 72 years, previously healthy and active, suspected of AMI*, within 12 hours. * typical symptoms, ECG S‐T elevation > 1 mm at least one lead or evolving pathological Q waves.   Exclusion criteria: Already receiving beta‐blockers or contraindications ( history of asthma, bracycardia < 60 bpm, AV block,  clinical or  radiographic evidence of cardiac failure. Attrition data: Screened: NR Total randomized patients: 100 Pindolol: n= 50 No pindolol (control) : n= 50   Total withdrawals (discontinuation of drug): NR   Withdrawals due to adverse events: NR   Total lost to follow‐up: NR   Baseline characteristics: Note: continuous variables are expressed as mean ± SD Mean age in years (range) Pindolol (P): 55.4 (28‐71) No pindolol or control (X): 54.4 (35‐71) Males, females P: 38, 12 X: 42,11 Site of infarction: inferior, anterior, other P: 28, 12, 10 X: 21,17, 12 Medical History Pindolol (P): smoking 40, Hypertension 13, IHD 12, previous AMI 4 No pindolol or control (X): smoking 42, Hypertension 13, IHD 16, previous AMI 9 SBP/DBP (mm Hg) Pindolol (P): 143.86 /89.44 ; 75.88 No pindolol or control (X): 140.20/ 89.93   ; 76.10
Interventions	Pindolol vs. No pindolol (control)    Drug regimen: Pindolol: 3 mg IV over 15 minutes every 8 hours x 3, followed by 5 mg orally every 8 hours for 6 doses No pindolol group or control group was manage identically as pindolol group (except for pindolol) with oxygen , anticuoagulation (15,000‐30,000 U/day) for five days, narcotic analgesics, sedatives, nitrate preparation, diuretic and anti‐arrythmic agents if necessary.   Other interventions: Pindolol group: morphine 22, nitrates 44 , diuretics 20, antiarrhythmics 26, DC cardioversion 2 Control group: morphine 34, nitrates 45 , diuretics 20, antiarrhythmics 24, DC cardioversion 3
Outcomes	Mortality:  obtained from table 3, page 708 Pindolol: 2day: NR ; 10day (hospital):1/50  ; eof : N/A   No pindolol or control (X): 2day:10day (hosp):1/50  ; eof : N/A Blood Pressure: Data was obtained from fig 1, page 707 . The mean BP change during first 24 hours  (20 h) was: Pindolol (n=50 ): SBP‐22.13 ±20.28 ; DBP‐9.92 ±12.06 Control (n=50 ): SBP‐13.06 ±22.92 ; DBP‐10.89 ±13.94 Standard deviation of the change was not reported but imputed from end point   Heart rate: Data was obtained from fig 1, page 707. The mean HR change during first 24 hours (20 h) was: Pindolol (n=50):‐3.33 ±13.38 Control (n=50):‐0.14 ±21.39 Standard deviation was not reported but imputed from end point
Notes	Funding: Not Reported (NR) Dates of conducting the trial : NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"random allocation", "Stratified according to timing of entrance (< 4 and > 4 hours)"
Allocation concealment?	Low risk	sealed envelope method
Blinding? All outcomes	High risk	Blinding was not addressed or stated
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods

Paci 1989.

Methods	Single‐site study (Italy) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 28 days Follow‐up: 28 days
Participants	41 patients within 12 hours after onset of symptoms of stroke Inclusion criteria: Sudden and persistent neurological deterioration due to a focal event in the carotid arterial distribution diagnosed as completed stroke, admitted to the hospital within 12 h after the onset of symptoms. Exclusion criteria: Transient ischemic attack, progressing stroke and cerebral hemorrhage were excluded. Patient unable to give informed consent or who have severe systemic disorders, recent myocardial infarction, congestive heart failure, abnormal hepatic pulmonary or renal functions or history of previous stroke. Attrition data: Screened: 54 (13 did not meet inclusion criteria, 4 TIA, 1 brain‐stem infarction, 1 progressing stroke, 5 as capsular hemorrhage). 1 was excluded ?glioma, 1‐dissection of aorta. Total randomized patients: 41, Nimodipine: n=  19, Placebo : n= 22 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: Total NR ( one patient witdrew due to skin rash‐ nimodipine group) Total lost to follow‐up: NR Baseline characteristics: Female: Nimodipine (N): 8, Placebo (P): 5 Age (years): N: 62 ±5, P: 63±6 Stroke localization (left hemisphere) N: 11, P: 13 SBP (mm Hg): N: 156.4 ±4.9, P:  149.3±5.5 DBP (mm Hg): N: 94.7 ±2.9, P:  85.9±2.4 Medical history (%) Nimodipine: diabetes 5, smoking 6, high‐fat 11, alcohol abuse 2, hyperlipidemia 5 Placebo: diabetes 3, smoking 10, high‐fat 9, alcohol abuse 1, hyperlipidemia 5
Interventions	Nimodipine vs.placebo Drug regimen: Nimodipine: 40 mg t.i.d orally for 28 days Placebo: identical aspect and duration No further details Other interventions: All patients received standard  treatment : 20% mannitol‐100 ml t.i.d for 7 days, nursing , physiotherapy, and ethically necessary drugs ( mostly anti‐hypertensive agents and antibiotics). Steroids, antiplatelet drugs, hyperosmolar agents, other calcium entry blockers and cerebral vasodilator were not administered
Outcomes	Mortality:  obtained from text, page 285 Nimodipine: 2day: 0; 10day:0  ; eof(28d‐N/A ):0  Placebo: 2day: 0; 10day:0  ; eof(28d‐ N/A ):0  Total non‐fatal SAE: NR Blood Pressure; Not reported during / up to the first 24 hours of treatment Heart rate: Not reported during / up to the first 24 hours of treatment
Notes	Funding: NR Dates of conducting the trial: NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"randomized" not further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Peter 1978.

Methods	Single‐site study (New Zealand) Open‐label Method of randomization: Not Reported Concealment of allocation: envelope method Duration of treatment: 27 hours Follow‐up: in Hospital
Participants	95 patients within 12 hours after onset of symptoms of AMI   Inclusion criteria: Moderate severity of AMI, within 12 hours of the onset of prolonged chest pain, ECG evidence of either epicardial injury ( > 2 mm ST elevation in anterior leads or > 1 mm in II, III, aVf, or pathological Q waves, and no contraindications.   Exclusion criteria: More than 65 years old, chest x‐rays or signs of pulmonary edema, less than 60 bpm, AV block > 1st  degree, had received BB within 72 hours, DC shock for ventricular arrhythmias, history of asthma.   Attrition data: Screened: NR Total randomized patients: 95 Propranolol: n=   47 no treatment (control) : n= 48   Note: cases were subdivided into three subgroups according to whether the pain started at 1) less than 4 hours, 2) 4‐8 hours, or 3) 8‐12 hours   Propranolol: n=   18, 17, 12 no treatment (control) : n= 19,17, 12 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR   Total lost to follow‐up: NR   Baseline characteristics: Age in years (range) Propranolol (P): 54 (37‐64) No treatment  or control (X): 54 (37‐64)   Position of infarct : anterior, Inferior, superior P: 17, 25, 5 X: 21,23,4
Interventions	Propranolol vs. open control Drug regimen: Propranolol: 0.1 mg/kg IV bolus over 10 minutes, followed by 320 mg orally over the next 27 hours* ( 40 mg at 1,3,7,11,15,19,23,27 hours )  Control : No specific treatment  * Treatment was stopped if cardiac failure or AV block developed or HR < 50 bpm  Other interventions: oral furosemide and lidocaine were allowed as necessary.
Outcomes	Mortality:  obtained from table 2, page 1092 Propranolol: 2day:NR ; 10day (hospital):1/47  ; eof (>30 d): NR no treatment (control) 2day:NR ; 10day (hospital):2/48  ; eof (>30 d): NR Blood Pressure: Data was obtained from figure 2, page 1093. The mean BP change during first 24 hours  (up 24) was: Propranolol (n=47): SBP ‐22.50±15; DBP‐13.64 ±14 Control (n=48): SBP‐7.50 ±17; DBP‐3.29 ± 14 Standard deviation of the change was not reported but imputed from end point   Heart rate: Data was obtained from , page . The mean HR change during first 24 hours was: Propranolol (n= 47): ‐10.61± 13.71 Control (n= 48):‐0.47 ±13.86 Standard deviation of the change was not reported but imputed from end point
Notes	Funding: Medical Researh Council and the National Heart Foundation of New Zealand
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"Patients were randomized" no further details
Allocation concealment?	Low risk	"by the enveloped method" comment: it is not describe it but it is given the benefit of doubt
Blinding? All outcomes	High risk	Blinding was not addressed or stated
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods

Pimenta 1985.

Methods	Single‐site study (Brazil) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 3 days Follow‐up: 22 months
Participants	20 patients within 6 hours after onset of symptoms of AMI Inclusion criteria: Clinic and ECG compatible with transmural AMI, aged < 75 years, free of heart failure, arrhythmias, A‐V block. Exclusion criteria: Receiving treatment with BB or other that might interfere with results Attrition data: Screened: NR, Total randomized patients: 20, Nifedipine, n= 10, Placebo, n= 10 Total withdrawals (discontinuation of drug): 2, Nifedipine: 2/10, Placebo: 0/10 Withdrawals due to adverse events: 2, Nifedipine: 2/10, Placebo: 0/10 Total lost of follow‐up: 0 Baseline characteristics: Age (years): Nifedipine (N): 57.5 ± 5.3, (51‐69, range), Placebo (P): 52.4±9.84, (39‐73, range) Males/femailes: N: 7/3, P: 7/3 Time to treatment  (hours) : V: 3.90 ±1.44, P: 3.44 ±1.13 SBP/ DBP (mm Hg): NR HR (bpm): NR Type of AMI: N: inferior 2, anterior 8, P: inferior 6, anterior 4
Interventions	Nifedipine vs. Placebo Drug regimen: Nifedipine: 10 mg / 6 hours  for 72 hours, followed by nifedipine 30 mg/ day Placebo: for 72 hours (no further details), followed by nifedipine 30 mg/ day Other interventions: All patients received opiaceos, oxygen (3‐5 l/min) benzodiazepine drugs
Outcomes	Mortality:  obtained it from text, page 11 Nifedipine: 2day: NR; 10day: 1/10 (acute phase); eof: (N/A)    Placebo: 2day:NR ; 10day: 0/10 (acute phase); eof: (N/A) Total non‐fatal SAE: NR Blood Pressure data during first 24 hours: NR Heart rate data during first 24 hours: NR
Notes	Funding: Not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients randomized" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not reported
Free of selective reporting?	Low risk	Mortality data reported according to methods

Pizzetti 2001.

Methods	Single‐site study (Italy) Open‐label Method of randomization: randomization block ( no further details) Concealment of allocation: patients were blindly assigned (no further details) Duration of treatment: 3 days Follow‐up: 6 months
Participants	90 patients within 3 hours after onset of symptoms of first AMI Inclusion criteria: Age < 70 years, diagnosis of AMI was based on ST segment elevation (> 2 mm in at least 2 leads, lasting > 30 min and not responsive to nitrates), dyskinesia on echocardiography  and increased  CK levels; no contraindication to thrombolytic therapy, admission within 3 hours of symptom onset.   Exclusion criteria: Previous myocardial infarction, CABG, heart failure or cardiogenic shock , bradycardia ( < 50 bpm) AV block , sick sinus syndrome,  < 100 SBP, and other cardiovascular or sever hepatic or renal disorders. Attrition data: Screened 281, 191 excluded, Total randomized patients: 90, Diltiazem (D) : n=43, Placebo (P): n=47 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost of follow‐up: NR Baseline characteristics: Age (years): Diltiazem (D):56±11, Placebo (P): 55 ±11 Sex (M/F): D: 27/16, P:39/8 BP (mm Hg) : NR Time to thrombolysis (min): D: 123 ±63; P: 144 ±56 Anterior infarction: D: (56 %) , P:  (51 %) Medical history (%) Diltiazem: hypertension 22(51) diabetes 14(38), Hypercholesterolemia 18 (42), smoking 28 (65) Placebo: hypertension 20 (43) diabetes 17(36), hypercholesterolemia 21 (43), smoking 35 (74)
Interventions	Diltiazem vs. Placebo Drug regimen: diltiazem: 0.15mg/kg bolus, followed by an infusion of 0.15 mg/kg/hour lasting 3 days. The maximum daily dose was 300 mg. Placebo: no further details Other interventions: All patients received intravenous nitrates, titrated according to blood pressure, aspirin ( 300 mg followed by 100 mg/day), recombinant tissue‐type plasminogen activator ( 100 mg; 15 mg bolus, 50 mg in 30 min, 35 mg in 60 min) heparin ( 5000 IU bolus, followed by infusion 1000IU/hour), propranolol  ( 1 mg iv bolus every 15 min to max 5 mg, then, atenolol oral 25‐100 mg/day according to target  < 140/90. Diltiazem group:  Beta‐blockers 6 (14%), nitrates 40(93%), heparin 43(100 %), ACEI 10 (23%), amiodarone 1(2%) Placebo group:   Beta‐blockers 13(28%), nitrates 42 (89%), heparin 47(100%), ACEI  10(21%), amiodarone 2(4 %)
Outcomes	Mortality:  obtained it from text and table IV, page 761 Diltiazem: 2day:NR ; 10day: 1/43 (during admission); eof(6 months ): 3/43  Placebo: 2day:NR ; 10day: 1/47(during admission)  ; eof(6 months): 2/47 Total non‐fatal SAE: NR Blood Pressure; Data was obtained from fig 1, page 761. The mean BP change during first 24 hours  (up to hour 24) was: Diltiazem (n=43 ): SBP ‐18.36 ± NR ; DBP NR Placebo (n=47 ): SBP ‐34.09 ± NR; DBP NR Heart rate Data was obtained from fig 1, page 761. The mean HR change during first 24 hours  (up to hour 24) was: diltiazem (n=43 ): ‐8.19 ±NR Placebo (n=47 ): ‐3.97 ±NR
Notes	Funding: Not stated, however it was acknowledged that Sanofi‐Synthelabo, Milan Italy supplied drug Dates of conducting the trial : March 1996‐Dec 1998
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"randomization in blocks", probably done adequately
Allocation concealment?	Low risk	"patients were blindly assigned"
Blinding? All outcomes	High risk	Blinding was not addressed or stated
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Free of selective reporting?	Low risk	Mortality data reported according to methods

PRACTICAL 1994.

Methods	Single‐site study (New Zealand) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 3 months Follow‐up: 12 months
Participants	225 patients within 24 hours after onset of symptoms of AMI   Inclusion criteria: Chest pain associated with either ST elevation in > 2 contiguous ECG leads, new pathologic Q waves or elevation on CK Exclusion criteria: Persistent hypotension with systolic blood pressure < 90 mm Hg, a history of sensitivity to ACE inhibitors  or the use of ACE inhibitors within 1 week of the AMI, valvular stenosis, severe renal or hepatic disorders, or a clear indication for treatment with an ACE inhibitor. Attrition data: Screened: 523, Excluded: 298 (101‐consent declined, 80‐administrative, 46‐taking ACEI, 41‐after 24 hours, 27 ‐hypotension, 2‐malignancy, 1‐known adverse effect to ACEi) Total randomized patients: 225, Captopril: n= 75, Enalapril: n= 75, Placebo: n=75 Total withdrawals (discontinuation of drug): 42, Captopril : 18, Enalapril: 12, Placebo: 12 Withdrawals due to adverse events: NR Captopril : total NR, (5‐hypotension, 3‐rash) Enalapril: total NR, (5‐hypotension, 4‐rash, ) Placebo: total NR, (2‐hypotension, 1‐rash) Total lost to follow‐up: NR Baseline characteristics: Age (years),  Captopril (C):64, Enalapril (E): 63, Placebo (P): 64 Men / women: C: 48/16; E: 59/ 16; P: 58/ 7 Thrombolytic therapy (%): C:51 (68),E:56 (75), P:55(73) Mean SBP (mm Hg): C: 133, E:139, P:129 Location and type C:  anterior 34(45), inferior 40(53),q wave 52(69),non‐q wave 23(31) E:  anterior 37(49), inferior 33(44),q wave 52(69),non‐q wave 22(29) P:  anterior 37(49), inferior 38(51),q wave 55(73),non‐q wave 19(25), Prior medical history (%) Captopril: AMI 13(17),CHF 2(3),CABG 5(7),PTCA 2(3), Enalapril: AMI 10(13),CHF 0(0),CABG 5(7),PTCA 1(1), Placebo: AMI 8(11),CHF 4(5),CABG 0(0),PTCA 1(1),
Interventions	Captopril vs. enalapril vs. Placebo Drug regimen: Captopril : 6.25 every 2 hours x  3 doses, followed by 25 mg 3 tid x 12 months Enalapril: 1.25 mg every 2 hours x 3 doses  followed by 5 mg tid x 12 months Placebo: same regimen, No further details Other interventions: All patients received thrombolytics, B‐blockers. Open‐label ACE might be given for a clear indication.
Outcomes	Mortality:  obtained from fig 5, page 1185 enalapril: 2day:NR ; 10day:1/75; eof (12mo‐N/A): 2 /75 captopril: 2day:NR ; 10day:6/75; eof (12mo‐N/A): 10 /75 Placebo: 2day:NR ; 10day:5/75; eof (12mo‐N/A): 12/75 Total non‐fatal SAE: NR Blood Pressure :  Not reported for the first 24 hours Heart rate : NR
Notes	Funding: Not stated but trialist acknowledged that Merck Sharp & Dohme (NZ) Ltd and Bristol‐Myers Squibb Ltd for supplying active drug and placebo Dates of conducting the trial : NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients randomized", no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind",
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Salathia 1985.

Methods	Single‐site study (UK) Double‐blind Method of randomization: in two separate blocks (mobile coronary and for other admissions) Concealment of allocation: NR Duration of treatment: 1 year Follow‐up: 1 year
Participants	800 patients within 6 hours after onset of symptoms of AMI   Inclusion criteria: Suspected AMI* within 6 hours of the onset. * The diagnosis of AMI was based on enzymes aspartate aminotransferase, lactic dehydrogenase and total creatin phospphokinase more than double.   Exclusion criteria: Delay from onset of pain exceeded 6 h, ventricular fibrillation, agonal rhythm, SBP < 90, HR > 100 bpm, clinical pulmonary oedema or congestive heart failure, bradycardia < 60 bpm, had received beta‐blockers within 48 hrs, AV block greater than 1 degree   Attrition data: Screened: 3350, Total randomized patients: 800, Metoprolol: =416, Placebo: n= 384 Total withdrawals (discontinuation of drug): NR Total withdrawals due to adverse events: NR Due to bronchospasm in hospital: Metoprolol: 3, Placebo: 1 Total lost to follow‐up: (at 1 year‐) Metoprolol: 3, Placebo : 1 Baseline characteristics: Age n (%): < 65 years, > 65 years Metoprolol: 287 (69%); 129 (31%) Placebo: 261(68%), 123(32%) Gender n (%), male, females Metoprolol (M): 290 (69.7%);126 (30.3%) Placebo (P): 282 (73.4%), 102 (26.6%) BP (mm Hg): NR Medical history n (%) Metoprolol: previous MI 105 (25.2), hypertension 50(12), smoking 194(46.6), angina 196(46.6), dyspnoea 117(28.1), Placebo: previous MI 109(28.4), hypertension 42(10.9), smoking 182(47.4), angina 176 (45.8), dyspnoea 110 (28.6),
Interventions	Metoprolol vs. placebo Drug regimen: Metoprolol: initial  15 mg IV  bolus over 5 min. Followed by oral metoprolol* 50 mg every 6 hors for 48 hours, and then 100 mg every 12 hours for one year Placebo: injection and tablets with similar appearance. No further details *tablets were stopped if: patients declined to continue, diagnosis was rejected, adverse reaction, interstitial or alveolar pulmonary oedema, AV block > 1 degree, SBP <80 mm Hg for more than an hour, persistent bronchospasm. Other interventions: Administration of other beta‐blockers and verapamil was avoided, no further details.
Outcomes	Mortality:  obtained from text/figure 1, page 193 Metoprolol: 2day: NR; 10day (hospital): 25/416; eof(1 year): N/A  49/416 Placebo: 2day: NR; 10day (hospital): 20/384 ; eof(1 year): N/A 52/384 Total non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR Heart rate change during first 24 hours: NR
Notes	Funding: Not stated, but authors acknowledged that Astra Pharmaceuticals supplied drug and placebo Dates of conducting the trial: NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"patients were randomized in two separate blocks (mobile coronary and for other admissions)", probably done
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Schulman 1995.

Methods	Single‐site study (US) Double‐blind Method of randomization: with stratification according to infarct location Concealment of allocation: NR Duration of treatment: 1 month Follow‐up: 1 month
Participants	43 patients within 24 hours after onset of symptoms of AMI   Inclusion criteria: BP ? 105/65. Diagnosis of AMI was based on evidence of ? 2 mV ST‐segment elevation or new pathologic Q waves in ? 2 contiguous ECG leads. Exclusion criteria: Renal (creatinine > 2.5 mg/dl), liver, neurologic or other life‐threatening diseases and treatment with (within 1 week) or known contraindications to ACE inhibitors; prior AMI, aortic stenosis, or cardiomyopathies.   Attrition data: Screened: NR , Total randomized patients: 43 , Enalapril : n= 22, Placebo : n= 21 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: Age (years): Enalapril (E): 58, Placebo (P): 63 Men/ women: E: 16/6, P: 11/10 Anterior myocardial infarction: E: 11, P: 12 SBP/DBP (mm Hg): E: 116/ 72, P:125/ 71 Time to treatment: E: 19.8 ± 1.3, P: 17.1±1.3
Interventions	Drug regimen: (same as that of CONSENSUS II trial) Enalaprilat: 1 mg IV over 2 to 3 hours*. Then, 6 hours after  oral  enalapril at dose of 2.5 mg bid titrated up to 20 mg /day at forth day and thereafter for 1 month.  Placebo:  No further details * The infusion was discontinued if systolic/ diastolic blood pressure decrease to < 100 / 60 mm Hg. mean dose at hospital discharge was 12 ± 2 mg /day   Other interventions: according to study group: Enalapril group : Thrombolytic therapy 19 (86%), IV nitroglycerin 19 (86 %), IV beta adrenergic blockers 7(32%), CCB 3 (14%) Placebo (P): Thrombolytic therapy 19 (90%), IV nitroglycerin 19 (90 %), IV beta adrenergic blockers 4 (19%), CCB 3 (14%)
Outcomes	Mortality:  obtained from text, page 766 Enalapril: 2day: NR; 10day: 2/22; eof (1 month ): N/A  Placebo: 2day: NR; 10day: 1/21; eof (1 month ): N/A   Total non‐fatal SAE: NR   Blood Pressure: Reported as mean BP (MAP) change from baseline to hour 6; as ‐5.8±2 in ACEI group, and ‐ 4.7 ± 2 mm Hg in placebo group. This data is not useful for the purpose of this review
Notes	Funding: Merck Sharp & Dohme Research Laboratories Dates of conducting the trial : NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"Randomization with stratification according to infarct location"
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Sirnes 1984.

Methods	Multi‐centre (4, Norway) Double‐blind Method of randomization: in blocks of 10 Concealment of allocation: given capsules from pre‐numbered bottles, no further details Duration of treatment: 6 weeks Follow‐up: 6 weeks
Participants	227 patients within 12 hours after onset of AMI Inclusion criteria: Severe chest pain for at least 30 min, ECG suggestion AMI ( not previously recognized ST elevation > 0.2 mV in precordial leads or > 0.1 mV in extremity leads or a Q wave ? 0.04 sec). Patient was considered included after taken the first capsule sublingually. Exclusion criteria: Age < 35, > 75 years, evaluation ? 12 h after onset, use of CCB within last 48 h, other serious disease, death before inclusion, refuse to participate. Attrition data: Screened 885, Total randomized patients: 227, Nifedipine : n= 112, Placebo : n= 115 Total withdrawals (discontinuation of drug): 5 protocol violation (2 nifedipine group, 3 placebo group) Withdrawals due to adverse events: Nifedipine : 9/110, Placebo : 6/112 Total lost of follow‐up: NR Baseline characteristics: Women (%): Nifedipine:23; Placebo: 29 Note: continuous variables are expressed as mean ± SD Age (years): Nifedipine: 61±8, Placebo: 61 ±9 SBP (mm Hg): Nifedipine: 147±30, Placebo: 144±27 DBP (mm Hg) :Nifedipine: 93±17, Placebo: 93±11 HR (bpm): Nifedipine: 75±18, Placebo: 77±22 Time to study entry (hours), Nifedipine: 5.4±3, Placebo: 4.9±2.9 Medical history (%) Nifedipine: previous AMI 28, angina 60, long‐term use of BB 18 Placebo: previous AMI 34, angina 53, long‐term use of BB 24
Interventions	Nifedipine vs. Placebo Drug regimen: Nifedipine: 10 mg capsule sublingually, subsequently 10 mg orally five times per day for 2 days, and 10 mg four times a day for 6 weeks. Treatment was temporarily discontinued if severe reaction developed. Placebo: no further details given. Other interventions: Other medical treatment was permitted according to local routine
Outcomes	Mortality:  obtained it from figure 1, page 640 Nifedipine: 2day:NR ; 10day: 6/112  ; eof(6 weeks): not applicable  Placebo: 2day:NR ; 10day:4/115  ; eof(6 weeks): not applicable Total non‐fatal SAE: NR Blood Pressure: (mm Hg) Data was obtained from table 5, page 642 The mean BP change during first 24 hours (up to hour ˜16) was: Nifedipine (n=110): SBP ‐17.67±22 ; DBP ‐9.67 ±14 Placebo (n=112 ): SBP ‐5.67 ±23 ; DBP ‐2.33±13
Notes	Funding: Bayer A.G. Leverkusen, Federal‐Republic of Germany Dates of conducting the trial : NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"patients randomly assigned in blocks of 10"; probably done as per concealment allocation
Allocation concealment?	Low risk	"given capsules from pre‐numbered bottles"
Blinding? All outcomes	Low risk	"double‐blind
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Theroux 1998.

Methods	Single‐site study (Canada) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 1 month Follow‐up: 1 month
Participants	60 patients within 6 hours after onset of AMI Inclusion criteria: Chest pain within previous 6 hours with ST‐elevation (1 mm ot motr in two or more adjacent leads), absence of a contraindication to thrombolysis or diltiazem and signed consent. Exclusion criteria: Not stated Attrition data: Screened: NR Total randomized patients: Diltiazem: n= 30, Placebo: n=30 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: Diltiazem: 2/30  (hypotension or bradycardia or supraventricular block) Placebo : 1/30 (developed shock before study drug was received) Total lost of follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years): Diltiazem (D):57±11, Placebo (P):60±9 Male: D: 25(83%), P:  24(83%) BP (mm Hg); NR Time of thrombolysis (minutes) : D: 175±64, P: 151±48 Site of AMI: D: anterior 12 (40), inferior 18(60) P: anterior 13 (45), inferior 16(55) Medical history (%) Diltiazem:  smokers 16(52), diabetes 2(6), hypertension 6(19),coronary artery disease 10(33) Placebo:  smokers 13(45), diabetes 3(10), hypertension 6(21),coronary artery disease 10(34)
Interventions	Diltiazem vs. Placebo Drug regimen: Diltiazem: 10 mg iv bolus in 10 min followed by an infusion at a rate of 10 mg/h for 48 hours. Oral therapy with diltiazem 120 mg three times a day (360 mg /day) until 4 weeks of follow‐up Placebo: no further details Other interventions: All patients receive tissue‐type plasminogen activator ( 7 mg iv bolus in 3 min followed by 53 mg in 57 min, 20 mg in 1 h and 10 mg/h for 2 h). Aspirin 325 mg orally at admission and daily thereafter. Heparin 1000 U/h with discontinuation after 72 hours. Diltiazem group: Nitroglycerin  19, BB 12, ACEi 6 Placebo group: Nitroglycerin  23, BB 22, ACEi 12 Other CCB were prohibited.
Outcomes	Mortality:  obtained from text, page 624 Diltiazem: 2day:0/30 ; 10day:0/30; eof(4 wks ): N/A  Placebo: 2day:1/30 ; 10day:2/30; eof(4 wks ): N/A Total non‐fatal SAE: NR Blood Pressure; Data was obtained from fig 2, page 624. The mean BP change during first 24 hours (up to hour 24) was: Diltiazem (n=30 ): SBP ‐16.26±16.21; DBP ‐8.02±11.64 Placebo (n=29): SBP ‐10.42±19.65; DBP ‐7.55±14.97 Heart rate Data was obtained from fig 2, page 624. The mean HR change during first 24 hours (up to hour 24) was: Diltiazem (n=30):‐3.73±11.61 Placebo (n=29 ): 1.50±14.65
Notes	Funding: Hoescht‐Marion‐Ruoussel
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"Patients were randomized"; no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

TIMI‐IIB 1991.

Methods	Multi‐centre (24‐US) Open‐label Method of randomization: unclear Concealment of allocation: NR Duration of treatment: 6 days (comparison), then same active treatment in both study groups for 1 year Follow‐up: 1 year
Participants	1434 patients within 4 hours after onset of symptoms of AMI   Inclusion criteria: Within 4 hours of the onset of AMI*, < 75 years, no contraindications to beta‐blockers. * AMI was based on typical chest pain ? 30 min, ST elevation 0.1 mV in two contiguous leads.   Exclusion criteria: Implanted pacemaker, resting ventricular rate < 55 bpm, SBP <100 mm Hg, moist rales or pulmonary edema + radiographic findings, advanced first‐degree or more advance heart block; asthma, chronic obstructive lung disease, b‐blocker, verapamil or diltiazem therapy on admission. Attrition data: Screened: 2,948, Excluded before randomization (n=1514) due to B‐blocker, verapamil or diltiazem therapy on admission (45.4%),  resting ventricular rate < 55 bpm (27.1%), SBP <100 mm Hg (21.3%), moist rales or pulmonary edema + radiographic findings (16.4%), advanced first‐degree  or more advance heart block (12.4%); asthma, chronic obstructive lung disease (11%), implanted pacemaker (0.3%), Total randomized patients: 1434, early metoprolol: n= 720 , No early metoprolol (deferred) group: n= 714 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Mean age (years): Metoprolol (M): 54.8 No early metoprolol or deferred group (X): 55.2 Race: white n (%): M: 614 (85.3), X: 626 (87.7) Sex: male n (%): M: 620 (86.1), X: 603 (84.4) BP (mm Hg): NR Medical History (%) Metoprolol group:  Prior AMI 47(6.5), angina 313(43.5), CHF 6(0.8), hypertension 216(30), diabetes mellitus 77(10.7), No early metoprolol group: Prior AMI 63(8.8), angina 313(43.8), CHF 10(1.4), hypertension 219(30.7), diabetes mellitus 82(11.5) Note: Prior myocardial infarction (overall): 7.7%
Interventions	Early Metoprolol vs. no early metoprolol (deferred  for first 6 days) Drug regimen: Early Metoprolol: Immediately after receiving rt‐PA, patients received 3 IV bolus of 5 mg at 2‐minute interval. Followed by, oral metoprolol at 25 mg every 12 hours for the first 24 hours and 100 mg every 12 hours thereafter*. No early metoprolol (deferred) group: Metoprolol started on day 6; 50 mg twice daily for 1 day and then 100 mg twice daily thereafter * Therapy was stopped if lengthening of PR interval beyond 0.26 sec, AV block 2 or 3 degree, wheezing or rales occurred. Therapy was temporary withheld if  HR <45 bpm, SBP < 90 mm Hg. Mean dose reported was not reported but percent of participants receiving B‐blockers: Early Metoprolol group: IV B‐blocker 651 (90 %), oral B‐blocker during first day 563 (78%) No early metoprolol (deferred) group: IV B‐blocker 33 (4.6%) oral B‐blocker earlier than scheduled (day 6): 22.9%  Other interventions: Most patients received rt‐PA intravenously ? 4 hours after onset of symptoms ( 87.6% in the early metoprolol group vs. 86.4% in the no early metoprolol or deferred group).  The total dose was 150 mg over 6 hours; however, because an unacceptable inicidenc of intracranial haemorrhage the dose was subsequently reduced to 100 mg in the remaining patients. Thus, initial 6 mg IV  bolus followed by 54 mg in the first hour , 20 mg in the second hour and 5 mg  in each of the next 4 hours. Patients also received lidocaine of a bolus of 1‐1.5 mg/kg followed by an infusion of 2‐4 mg/ min for 24 hours. Heparin 5,00 unit bolus followed by 1000 units/hr , dose adjusted to maintain an activated partial thromboplastin time of  1.5‐2  times control values, then 10,000 IU SC twice daily until hospital discharge. Aspirin 80 mg/day started on day 1 in 93 of patients (6.5%) or day 2 (93.5%). This dose was increased to 325 mg/ day on day 6.
Outcomes	Mortality:  obtained from table 6, page 430 Early metoprolol group: (n=720) At 6 days: 17; in hospital 22; at week 6: 26; at year 1: 34   No early  metoprolol (deferred) group: (n=714) At 6 days:17; in hospital 25; at week 6: 25; at year 1: 35   Total non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR Heart rate change during first 24 hours: NR
Notes	Funding: National Heart, Lung and Blood Institute, National Institutes of Health
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"patients randomized" in the circulation 1991;83:422‐437. "patients were assigned to treatment on the basis of randomization schedules designed to provide balance... within each clinical center.." N Engl J Med 1989;320:618‐627. Comment: probably done
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	High risk	open control
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Free of selective reporting?	Low risk	Mortality data reported according to methods

Tonkin 1981.

Methods	Single‐site study (1‐Australia) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 7 days Follow‐up: hospital
Participants	89 patients within 24 hours after onset of symptoms of first AMI   Inclusion criteria: Suspected diagnosis of first AMI within 24 hours of onset     Exclusion criteria: Receiving Beta‐blockers or having contraindications to these agents (sinus bradycardia, <45 bpm, SBP < 100 mm hg), P‐R interval greater than 0.22 s, 2 or 3 degree AV block, moderate or severe heart failure, and obstructive airways disease. Attrition data: Screened: NR, Total randomized patients: 88, Timolol: n= 42, Placebo: n= 46 Total withdrawals (discontinuation of drug): 36 Timolol: 6 (due to diagnosis) +11 (due to contraindication) =17   Placebo: 8 (due to diagnosis) +11(due to contraindication) =19 Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: 72 of patients were male, Mean time delay to treatment (hours) Timolol: 11.94 ± 4.65, P: 10.09 ± 5.44 The rest of baseline characteristics are not reported
Interventions	Timolol vs. placebo Drug regimen: Timolol: 10 mg twice / day for 7 days Placebo: No further details Other interventions: All patients were managed in a routine way, though IM injections were avoided. Anti‐arrhythmic and anti‐failure therapy were used as indicated
Outcomes	Mortality:  obtained from text, page 145 Timolol: 2day: NR; 10day (in hospital): 1/42; eof (1 year): NR  Placebo: 2day: NR ; 10day (in hospital):1/46; eof (1year): NR Total non‐fatal SAE: NR Blood Pressure change during first 24 hours: NR Heart rate change during first 24 hours: NR
Notes	Funding: Merck, Sharp and Dohme
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomized, no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	not reported
Incomplete outcome data addressed? 10‐day mortality	Unclear risk	it is not clear whether all randomized patiens were included in the mortality data as the statement "patients that completed study were:Timolol 25, Placebo 27"
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	not reported
Free of selective reporting?	High risk	Except for time to treatment and percent of male patients. Baseline characteristics are not reported

Van‐de 1993.

Methods	Multi‐centre (20‐Nehterlands ) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 10‐14 days Follow‐up: Hospital
Participants	201 patients within 5 hours after onset of symptoms of AMI   Inclusion criteria: Patients < 71 years of age, pain suggestive of acute MI*, onset of chest pain < 5 hours before initiation of therapy, no previous history of angioplasty or bypass surgery, no contraindications to thrombolytic therapy or to BB, no beta‐blocker or CCB within 1 week,.   *Diagnosis of AMI was based on pain lasting ? 30 min, ST elevation of 0.2 mV in two or more limbs leads, or leads V5‐V6 or 0.3 mV in two or more precordial leads  (V1 to V4) or ST elevation of 0.1 mV in two leads (II, III, AVF, or V5 and V6)  associated with ST depression of 0.2 mV Iin two precordial leads.   Exclusion criteria: Contraindications:  HR < 50 bpm, SBP < 90 mm Hg, CHF, shock, 2,3 AV block, bronchospasm, sick sinus syndrome   Attrition data: Screened: NR, Total randomized patients: 300, Atenolol: n= 103, Placebo: n=98 Note: another randomized group was alinidine, a derivative of clonidine, (n=99) which is not considered or discussed any further in this review Total withdrawals (discontinuation of drug):8 , Atenolol: 3, Placebo 4, Alinidine 1 Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: Mean age, in years,(range) : Atenolol (A): 59 (39‐70), Placebo (P): 57 (40‐68) Male: A: 80(80%), P:  80(85%) Location of the infarction: anterior, inferior, Not reported A: 33%, 66%,1%; P:  36%, 62%,2%, Time to treatment in minutes (range): A: 156 (85‐300), P:  168(72‐330) HR in bpm (range) : A: 74 (55‐111), P:  75(54‐102) SBP in mm Hg (range): A: 130 (100‐170), P:  135 (100‐170) Medical history (%) Atenolol: Previous MI 3(3), diabetes Mellitus 4(4), angina > 4 weeks 15(15), Placebo: Previous MI 3(3), diabetes Mellitus 5(5), angina > 4 weeks 10(11),
Interventions	Atenolol vs. Placebo Drug regimen: Atenolol: 5 mg IV twice with a 10 min interval 10 mcg /min, followed by oral treatment 25 to 50 mg of atenolol every 12 hours for 10‐14 days* *Dose was adjusted according to hemodynamic; 70 received full intravenous dose Placebo: two IV injections; no further details. It is not known if oral placebo was given.   Other interventions: All patients received Alteplase 100 mg over 3 hours, and heparin 5000 IU bolus, followed by continuous infusion of 1,000 IU /h until angiography. Aspirin was not given
Outcomes	Mortality: obtained from text in page 412 Atenolol : 2day:NR ; 10day:1/103  ; eof (hospital ): 1 Placebo: 2day:NR ; 10day:2/98  ; eof(hospital): 4 Total non‐fatal SAE: NR Blood Pressure: not reported during the first 24 hours Heart rate: not reported during the first 24 hours
Notes	Funding: Belgian National Fund for Scientific Research, ICI Pharma, Belgium and Boehringer Ingelheim, Belgium. Dates of conducting the trial : June 1988‐ December 1990
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomized", no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"use of a double‐dummy technique"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

VENUS 2001.

Methods	Multi‐centre ( Netherlands) Double‐blind Method of Randomization: Simple, equal blocks of 10 according to computer‐generated lists. Concealment of allocation: Numbered boxes contained one complete treatment or identical placebo course and were sequentially distributed among participants Duration of treatment: 10 days Follow‐up: 3 months
Participants	454 patients within 6 hours after onset of symptoms of acute stroke Inclusion criteria: Patients with acute stroke and hemiparesis Exclusion criteria: Ability to raise arm or leg > 10 seconds against gravity, inability to start treatment within 6 hours, age <18 or > 85 years, previous participation in this trial, pregnancy, impaired consciousness ( did not obey orders and did not open eyes on painful stimuli); other diseases likely to cause death within 1 year,  previous stroke, resulting in serious handicap, dysphagia, excluding oral medication at trial onset; systolic blood pressure < 130 mm Hg, heart rate < 50 bpm; and ? 3 of the following 4 conditions: severe headache, vomiting, hypertension ( SBP > 220 mmHg) and use of oral anticoagulants. Attrition data: screened / excluded at screened: NR Total randomized patients: 454, Nimodipine: n= 225, Placebo : n= 229 Total withdrawals (discontinuation of drug): 15,Nimodipine: 7/225, Placebo : 8/229 Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: Female: Nimodipine (N): 97(43%), Placebo (P):  85 (37%) Age, median (range): N: 70.5 (24‐91), P:  71.1 (31‐93) Medical history Nimodipine (N): previous stroke 7(3%), cardiac disease 46(20%), other 105(47%)  Placebo (P): previous stroke 17(7%)*, cardiac disease 60(26%), other 108(47%) Neurological examination N: Hemiparesis 189(84%), impaired consciousness 1(0%), aphasia 66(29%)  P: Hemiparesis 187(82%), impaired consciousness 3(1%), aphasia 89(39%)* Type of stroke N: Ischemic 133(59), hemorrhagic 20 (9), no CT scan 9(32), no stroke 1(0), P: Ischemic 128(56), hemorrhagic 15(7),no CT scan 79(34), no stroke 7(3)
Interventions	Nimodipine vs. Placebo Drug regimen: Nimodipine: 30 mg orally every 6 hours for 10 days Placebo: No further details
Outcomes	Mortality:  obtained text and table 2, page 463 Nimodipine: 2day: NR; 10day: 14/225(6%); eof (3 mo): NR*  Placebo: 2day:NR ; 10day:20/229(9%) ; eof (3mo):NR* *Mortality alone was NR at 3 months but a composite of all‐cause mortality or dependency in life ( Modified ranking scale score > 3)  as 32% vs. 27%, RR, 1.2; 95% CI 0.9‐1.6, p value NS. Total non‐fatal SAE: NR Blood Pressure on first 24 hours : NR Heart rate on first 24 hours: NR
Notes	Funding: Dutch Prevention Fund (grant 28‐2467). Trial medication was provided by Bayer AG, Germany Dates of conducting the trial : ended on July 1998
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Simple, equal blocks of 10 according to computer‐generated lists
Allocation concealment?	Low risk	Numbered boxes contained one complete treatment or identical placebo course and were sequentially distributed among participants
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	not reported
Free of selective reporting?	High risk	Mortality was not reported long‐term when the methods stated it would be measured

von Essen 1982.

Methods	Two‐site study (41 in Germany; 15 in Switzerland) Double‐blind Method of randomization: according to a randomization list, no further details Concealment of allocation: NR Duration of treatment: 14 Follow‐up: 14
Participants	51 patients within 24 hours after onset of symptoms of AMI   Inclusion criteria: Acute chest pain typical of infarct, interval between acute symptoms and hospitalization ? 24h, symptom of infarct  with persistent pain, SBP > 110mm Hg, heart pulse = 60 bpm, CI 4.0 L/min/ m2   Attrition data: Screened: NR, total randomized patients: 51, Metoprolol: n= 25, Placebo : n= 26 Total withdrawals (discontinuation of drug): NR Metoprolol : 1 (due to bradycardia) Withdrawals due to adverse events: Metoprolol : 1 /  25 Placebo : NR Total lost to follow‐up: NR Baseline characteristics: Age (years): between 38‐83 years old, overall mean 59, 45 men, 6 women BP (mm Hg): NR Time to treatment (hours: minutes) Metoprolol : 9:55± 5 Placebo :  12± 6 History of previous acute myocardial infarction; 5/51 (10%) Metoprolol : 1, Placebo : 4 History of diabetes 8/51(16%) , Metoprolol : 3, Placebo : 5 History of hypertension 18/51 (35%), Metoprolol : 8, Placebo : 10
Interventions	Metoprolol vs. placebo Drug regimen: Metoprolol: Initial 0.1 mg/kg body weight rapid IV infusion, 2 hours later 100 mg and then 100 mg every 12 hours Placebo: No further details
Outcomes	Mortality:  obtained from text, page 1267 & 1271 Metoprolol: at 2day:NR; 10day: 1*/ 25; end of follow‐up:   (14d) 1/25 Placebo: at 2day:NR; 10day: 1*/ 26 ; end of follow‐up: (14d) 1/26 * It is assumed that these deaths occurred during the first 10 days, as it is stated that these deaths occurred right after the acute phase of 48 hours. Total non‐fatal SAE: NR Blood pressure: NR Heart rate: NR
Notes	Funding: Ciba‐Geigy
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"patients allocated according to a randomization list", probably done
Allocation concealment?	Low risk	according to a randomization list, probably done
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Wagner 2002.

Methods	Single‐site study (Austria) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 1 day Follow‐up: 7 days
Participants	99 patients within 6 * hours after onset of symptoms of AMI with undergoing thrombolysis   * this is assumed value as in clinical practice ( in 2001) patients are usually treated with thrombolysis within 6 hours. In addition, the mean time of chest pain was 2.1 hours.  Inclusion criteria: Diagnosis of AMI was based on chest pain lasting for more than 30 min, ST‐elevation > 1 mm in one or more inferior leads or in at least two corresponding anterior leads and a typical rise and fall in Ck and CK‐MB. Q‐wave infarction was confirmed by serial electrocardiographic abnormalities, with development of Q‐waves as well as typical rise and fall or CK‐MB. Exclusion criteria: Chest pain relieved by nitroglycerin or < 30 min in duration; 2) history of a myocardial infarction; contraindications to thrombolytic therapy or ramipril, serious advanced illness, hypotension SBP < 100 mm Hg, cardiogenic shock, use of ACE within 2 weeks, pregnancy, lactation, or inability to participate.    Attrition data: Screened: NR, total randomized patients: 99, ramipril : n= 51, placebo : n= 48 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age (years): Ramipril (R): 55 ± 12, Placebo (P): 55±11 Male/female: R:79/21, P:69/31 Anterior / inferior MI: R : 25 /26, P : 22 / 26 Duration of chest pain (hours) : R: 2.1 ± 1.6, P:  2.1 ± 1.7 SBP/DBP (mm Hg): R: 139 ±19 / 77±13, P: 138 ±20 / 77±15 Heart rate: R: 80 ± 20, P:  79 ± 17 Medical history (%) Ramipril: hypertension (43), diabetes (14), smoking (48), hyperlipidemia (34) Placebo: hypertension (29), diabetes (16), smoking (52), hyperlipidemia (36)
Interventions	Ramipril vs. placebo Drug regimen: Ramipril: 2.5 mg orally prior to thrombolysis (rt‐PA). Then, a second dose of 2.5 mg, 12 hours after. Placebo: prior to thrombolysis and then a second dose 12 hours after, no further details. 24 hours the start of thrombolysis all patients received ramipril with a starting dose of 2.5 mg   Other interventions: All patients received 100 mg rt‐PA according to the GUSTO‐scheme 5000 IU heparin and 100 mg aspirin orally. Beta blockers were given as clinically appropriate.   Ramipril  group: aspirin 10 (%), beta blockers 24 (%), CCB 1(%), diuretics 3 ( %), nitrates 1 (%) Placebo  group: aspirin 14(%), beta blockers 22(%), CCB 2(%), diuretics 4( %), nitrates 1(%)
Outcomes	Mortality:  obtained from text, page 183 Ramipril: 2day: NR ; 10 day (within 1 week) 1/51 :  eof : N/A Placebo: 2day: NR ; 10 day (within 1 week) 1/48 : eof : N/A   The causes of death were rupture of the anterior wall or the left ventricle in one patient and cardiogenic shock in the other one.   Total non‐fatal SAE: NR Blood Pressure; Data was obtained from text in page 183. Reported as mean BP at end point ( 24 h) : Ramipril (n=51 ): SBP 108 ± 15 ; DBP 72 ± 12 Placebo (n=48 ): SBP 112 ±14 ; DBP 73 ± 10   Heart rate Data was obtained from text in page 183. Reported as mean HR at end point ( 24 h) : Ramipril  (n= 51 ): 70 ± 12 Placebo (n= 48): 73 ± 11
Notes	Funding: Aventis Pharma (provided study medication) Dates of conducting the trial :  January 1999‐February  2001
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomly assigned" no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Yusuf 1983.

Methods	Single‐site study (UK) Open‐label Method of randomization:  NR Concealment of allocation:  using numbered, sealed envelopes. Duration of treatment: 10 days Follow‐up: 1‐ 4 years (Average 2)
Participants	477 patients within 12 hours after onset of symptoms of suspected AMI   Inclusion criteria: Suspected of AMI* within 12 hours of onset.   *According to ECG patients were subdivided into those with definite MI (i.e., ST‐ segment elevation of at least 1 mm limbs leads, or 2 mm in precordial leads (with or without Q waves),  and those with threatened MI ( all other who were again subdivided into those with suggestive ? T‐wave inversion, ST depression or BBB‐ or those with no particular abnormality on ECG)   Exclusion criteria: HR <40 bpm, SBP < 90 mm Hg, > second degree AV block, heart failure requiring digoxin or furosemide 80 mg , or history of asthma, taking BB at entry, or thought requiring BB at admission, have contra‐indication. Attrition data: Screened: NR, Total randomized patients: 477, Atenolol: n= 244, No atenolol or control: n= 233 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost to follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD (if provided) Mean age ‐ in years: Atenolol (A): 56, No atenolol or control (X): 56 Female (%): A: 30 (12), X: 43 (18) BP (mm Hg): A: 143 / 89, X:  145 / 91 HR (bpm): A: 77, X: 77 Time to randomization (hours): A: 5, X: 5 Heart failure at entry (%): A: 15 (6), X: 16 (7) Site of myocardial infarction: anterior, inferior, both, indefinite (%) A: 100(41),86(35), 12(5), 46(19). X: 91(39), 86(37), 5(2), 51(22). Medical history (%) Atenolol group: Angina 25(10), hypertension 29 (12), myocardial infarction 44(18), diabetes 14(6) Control group: Angina 29(12), hypertension 28 (12), myocardial infarction 34(15), diabetes 9(4),
Interventions	Atenolol vs. no treatment (control) Drug regimen: Atenolol: 5 mg IV over 5 min, then 50 mg, orally, immediately after and 12 hours later. Then, 100 mg once daily for 10 days or until the patient develop contraindications died or was discharged.   Control: BB was not allowed unless clearly indication Other interventions: All patients received routine ancillary management Atenolol group: n=244 atropine 23, inotropic 4, antiarrhythmic 58, other anti‐hypertensive 14, CCB 9, oral anticoagulants 53,antiplatelets drugs 13, bronchodilator 3, intra‐aortic balloom pump 0    No atenolol or control: n= 233 atropine 19, inotropic 6, antiarrhythmic 76*, other anti‐hypertensive 41*, CCB 22*, oral anticoagulants 48,antiplatelets drugs 15, bronchodilator 4, intra‐aortic balloom pump 2.             * statistically significantly different
Outcomes	Mortality:  obtained from table 5 , page I‐38 Atenolol: n= 244  2day: NR; 10day:7; eof (average 2 years ): 36 No atenolol or control: n= 233 2day: NR; 10day:16; eof (average 2 years ): 44 Total non‐fatal SAE: NR Blood Pressure: Not reported change during first 24 hours of all randomized patients Heart rate: Not reported change during first 24 hours of all randomized patients
Notes	Funding: British Heart Foundation and ICI Pharmaceuticals Dates of conducting the trial : August 1978‐ May 1981
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"patients randomized"; probably done as description in allocation concealment
Allocation concealment?	Low risk	Because a 24‐hour telephone randomization service was not available, randomization was done using numbered, sealed envelopes. To prevent unauthorized withdrawals, and to ensure unbiased completeness of certain initial data, the patient’s name heart rate and blood pressure should be written on the outside of the envelope before it was opened. Once a sealed envelope had been opened for a new  patients, that patient was irrevocably entered, even if the initial criteria had not been fully met.
Blinding? All outcomes	High risk	Blinding was not stated
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	not reported
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Free of selective reporting?	Low risk	Mortality data reported according to methods

Zannad 1988.

Methods	Single‐site study (France) Double‐blind Method of randomization: NR Concealment of allocation: NR Duration of treatment: 3 weeks Follow‐up: 3 weeks
Participants	34 patients with clinical, ECG and enzymatic evidence of AMI within 6 hours after onset of symptoms of AMI Inclusion criteria: Chest pain, epicardial injury and pathological q waves, less than 6 hours since the onset of symptoms, no previous AMI, heart failure (KK II), or 3rd degree AV block, no ongoing use of amiodarone, BB or CCB, age < 75, ability to provide informed consent. Exclusion criteria: Not stated Attrition data: screened: NR Total randomized patients: 34, Diltiazem: n= 17, Placebo: n= 17 Total withdrawals (discontinuation of drug): Diltiazem:  0/17, Placebo: 2/17 Withdrawals due to adverse events: NR Total lost of follow‐up: NR Baseline characteristics: Note: continuous variables are expressed as mean ± SD Age‐ in years: Diltiazem (D):54 ±10, Placebo (P): 49 ±10 Male/female: D:  14/3, P: 14/3 Time of infarction (hours): Diltiazem (D):4.6 ±0.9, Placebo (P): 4.4 ±1.0 Inferior/ Anterior involvement : D:  11/6, P: 11/6
Interventions	Diltiazem vs. Placebo: Drug regimen: Diltiazem: 10 mg iv injection over 2 minutes, followed by a constant rate infusion of 15‐20 mg / h over 72 hours. (adjusted if side effects). Oral treatment started 2 hours before infusion was discontinued ; as 60 mg every 6 hours for 3 weeks.   Placebo: No further details mean dose reported was not reported Other interventions: Standard treatment: all patients received  anticoagulant doses of heparin and a constant rate infusion of 800 to 800 mg/ 24 of lidocaine. The use of nitrates, betablocker, amiodarone or ccb was not permitted during study.
Outcomes	Mortality:  obtained from text, page 1174 Diltiazem: 2day:1/17 ; 10day:1/17  ; eof(3 wks ):Not applicable  Placebo: 2day:0/17 ; 10day: 1/17 ; eof(3 wks ):Not applicable Total non‐fatal SAE: NR Blood Pressure data: NR Heart rate data : NR
Notes	Funding: Not Reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomly assigned"; no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	Low risk	"double‐blind"
Incomplete outcome data addressed? 2‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 10‐day mortality	Low risk	Mortality data was provided according to all randomized patients
Incomplete outcome data addressed? 30‐day mortality	Unclear risk	Not applicable as treatment lasted more than 10 days
Free of selective reporting?	Low risk	Mortality data reported according to methods

Zharov 1991.

Methods	Single‐site study (Moscou USSR) Open‐label Method of randomization: NR Concealment of allocation: NR Duration of treatment: 3 days Follow‐up: 25 days
Participants	115 patients with AMI within 12 h of onset Inclusion criteria: NR Exclusion criteria: NR Attrition data for the three randomized groups: Control (I ): n= 52 Isosorbide dinitrate (II): n=32 Nitroglycerin (III): n=31 Total withdrawals (discontinuation of drug): NR Withdrawals due to adverse events: NR Total lost of follow‐up: NR Baseline characteristics were not reported according to group: There were 87 males, 28 females, mean age 61± 0.9, primary AMI 70, repeated infarction 37 cases; 68 anterior, 23 anterolateral, 11 posterolateral, 13 posterior.
Interventions	Control (I ): n= 52 Isosorbide dinitrate (II): n=32 Nitroglycerin (III): n=31 Drug regimen: I: receive only heparin (no further details). II and III: the nitrate infusion were given at an initial rate of 25 mcg/ min and gradually titrated up until systolic blood pressure fell 12‐25 % of initial value, but not below 90 mm Hg. Other interventions: NR
Outcomes	Results based on following group population: Control (I ): n= 52 Isosorbide dinitrate (II): n=32 Nitroglycerin (III): n=31 Mortality: obtained it from table 3, page 67 Control : 2day:2 ; 10day:NR ; eof(25day ): 16 Isosorbide : 2day:NR ; 10day:NR ; eof(25day ):2 Nitroglycerin: 2day:NR ; 10day:NR ; eof(25 day): 3 Non‐fatal SAE: NR Individual SAE: CHF: obtained it from table 3, page 67 control:13/52 (25%) isosorbide: 5/32(16%) nitroglycerin: 4/31(13%) Aneurysm: obtained it from table 3, page 67 control:11/52 (21%) isosorbide: 7/32(22%) nitroglycerin: 6/31(19%) Pulmonary edema: (data on this outcome was confusing as different numbers are given in table 1, 2 and 3). Blood Pressure; reported as change in BP (mm Hg) from baseline for the two nitrates groups as follow: Isosorbide (n=32): SBP ‐14.7±3.8, DBP‐7.7±1.6 (text on page 65) Nitroglycerin (n=31): SBP:‐15.2±4.9, DBP ‐11.7±3.6 (text on page 67) BP data for the control group was not reported. Heart rate: reported as change in HR (bpm) from baseline for the one nitrate groups as follow: Nitroglycerin (n=31): ‐17.5±4.9 (text on page 67) BP data for the other two groups was not reported.
Notes	Funding: Not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"patients were randomly assigned", no further details
Allocation concealment?	Unclear risk	not reported
Blinding? All outcomes	High risk	Open‐control trial
Incomplete outcome data addressed? 2‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 10‐day mortality	Unclear risk	Not reported
Incomplete outcome data addressed? 30‐day mortality	Low risk	Mortality data was provided according to randomized patients
Free of selective reporting?	High risk	Incomplete information on baseline characteristics, withdrawals

DBP: diastolic blood pressure 
 EOF: end of follow‐up 
 ITT: intention to treat 
 NR= not reported 
 N/A= not applicable 
 SAE: serious adverse events 
 SBP: systolic blood pressure

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Annane 1996	Follow‐up less than 24 hours. A secondary reason for exclusion was that this trial only included responders to another anti‐hypertensive before entering to the trial
Ardissino 1997	Uncertain time of inclusion of patients
Azancot 1982	Mortality not reported
Azcona 1990	Mortality not reported. No publication was found for the results of the trial.
Balcon 1966	Mortality not reported according to our time‐frame. A secondary reason for exclusion was that results are not given on an intention to treat basis.
Barber 1976	Uncertain time of inclusion of patients. A secondary reason for exclusion was that results are not given on an intention to treat basis.
Basu 1997	Mortality not reported according to our time‐frame.
Blanc 1989	Mortality not reported.
Briant 1970	Uncertain time of inclusion of patients
Bussmann 1984	Mortality not reported.
CATS 1994	Mortality not reported according to our time‐frame. Letter sent to trialists on July 19,2008, trying to obtain data. We have not got a response yet
CHHIPS 2005	Mortality not reported. Publication of the trial results were never found published
Davalos 1992	Mortality not reported according to our time‐frame
DAVIT I 1984	Uncertain time of inclusion of patients. For example, It is stated that patients were to be included "on admission", but treatment started after 24 hours in >15% of patients with AMI and mortality data was not reported separately for those starting earlier or later than the 24 hours. An additional problem was that these AMI patients were part of a subset of the entire randomized patients (AMI‐41% + suspected AMI‐59%) based on intention to treat analysis. Therefore, it should not be dismissed the possibility of the inclusion of additional patients being admitted later than 24 hours if all randomized patients were considered
Emanuelsson 1984	Mortality was not reported
EMIP 1994	Mortality not reported according to our time‐frame.
Evemy 1978	Uncertain time of inclusion of patients
Eveson 2007	Mortality results are not given on an intention to treat basis. Letter sent to trialists on March 4, 2008, trying to obtain data. We have not got a response yet
FAMIS 1998	Mortality not reported according to our time‐frame. Sent letter to trialists on June 20,2008, trying to obtain data. We've got no response yet
Franke 1996	Mortality not reported according to our time‐frame. Sent letter to trialists on August 12,2008, trying to obtain data. We have not got a response yet.
French 1999	Mortality not reported according to our time‐frame. Sent letter to trialists on July 1, 2008, trying to obtain data. We have not got a response yet.
Gardtman 1999	Uncertain time of inclusion of patients
Gebalska 2000	Mortality not reported
Gelmers 1984	Uncertain time of inclusion of patients
Gerstenblith 1982	Uncertain time of inclusion of patients. Chronic Unstable angina
GISSI‐3p 1994	Mortality not reported according to our time‐frame.
Gonzalez‐Fernandez 1993	Uncertain time of inclusion of patients
Gordon 1984	Mortality not reported
Gottlieb 1986	Mortality not reported according to our time‐frame.
Gottlieb 1988	Mortality not reported according to our time‐frame. Letter sent to trialists on August 13, 2008, trying to obtain data. We have not got a response yet
Gupta 1984	Incomplete reporting‐ only abstract
Gupta 1985	Mortality not reported according to our time‐frame.
Hamilton 1996	Follow‐up less than 24 hours.
Haude 1991	Follow‐up less than 24 hours.
HEART 1997	Mortality not reported. Letter sent to trialists on August 4, 2008, trying to obtain data. We have not got a response yet
HINT 1986	Uncertain time of inclusion of patients, A secondary reason for exclusion was that results are not given on an intention to treat basis
Jaffe 1987	Mortality not reported
Just 1986	Mortality not reported
Kahler 1995	Mortality not reported according to our time‐frame.
Karlberg 1998	Mortality not reported according to our time‐frame.
Kolettis 1983	Mortality not reported according to our time‐frame.
Kumada 1995	Uncertain time of inclusion of patients
Lejemtel 1993	Mortality not reported. Letter sent to trialists on July 20,2008, trying to obtain data. We have not got a response yet. Publication of the trial results were never published
Lloyd 1988	Mortality not reported
Loogna 1985	Mortality not reported
Macleod 1980	Mortality not reported. There was only an abstract reported
Manolis 1999	Uncertain time of inclusion of patients. It could be argued that it was within 24 hours since study treatment was given right after thrombolysis, but we could not confirm this
Matias‐Gutierrez 1992	Mortality not reported according to our time‐frame.
McGrath 1986	Mortality not reported.
Morris 1995	Mortality not reported according to our time‐frame. Letter sent to trialists on August 13,2008, trying to obtain data. We have not got a response yet
Mueller 1980	Uncertain time of inclusion of patients
Murdock 1990	Uncertain time of inclusion of patients
Oldroyd 1991	Mortality not reported according to our time‐frame. Letter sent to trialists on July 24, 2008, trying to obtain data. We have not got a response yet.
Oshima 1997	Mortality not reported according to our time‐frame.
Osuna 1985	Uncertain time of inclusion of patients
Ramsdale 1988	Mortality not reported
Reinert 2004	Uncertain time of inclusion of patients
Renard 1987	Follow‐up less than 24 hours
Reynolds 1972	Uncertain time of inclusion of patients. A secondary reason for exclusion was that results were not given on intention to treat basis.
Schrader 2003	Mortality not reported according to our time‐frame. This trial included patients within 24h and within 36 h, but results are not reported separately for those two categories. Letter sent to trialists on May 1,2008, trying to obtain data. We have not got a response yet
Sloman 1967	Uncertain time of inclusion of patients
SMILE 1995	Mortality not reported according to our time‐frame. Letter sent to trialists on June 20,2008, trying to obtain data. We have not got a response yet.
Szczechowski 1994	Mortality results are not given on an intention to treat basis
Waagstein 1976	Follow‐up less than 24 hours
Walker 1988	Mortality results are not given on an intention to treat basis. Letter sent to trialists on August 13, 2008, trying to obtain data. We have not got a response yet
Wilcox 1980a	Mortality not reported according to our time‐frame.
Wilcox 1980b	Mortality not reported according to our time‐frame
Wilcox 1986	Mortality not reported according to our time‐frame. Last letter sent to trialists on August 13, 2008, trying to obtain data. We have not got a response yet
Wimalaratna 1994	Mortality not reported. Letter sent to trialists on August 13, 2008, trying to obtain data. We have not got a response yet
Zochowski 1986	Uncertain time of inclusion of patients

Differences between protocol and review

Changes from the protocol to the review:

Title

• The term "health outcomes" was changed to a more specific term "Mortality"

Objectives

• From "early treatment" changed to more specific terms "immediate and short‐term treatment"

• From "short‐term and long‐term health outcomes" to more specific  terms "mortality at 2,10 and ≥ 30 days"

• Regarding BP and HR changes from "during time the drug is being administrated" to more specific terms "during the first 24 hours of treatment"

Types of studies

• From "RCT must provide at least one of primary outcome measures" to more specific terms "must provide mortality data at least one of the specified time periods, 2 days, 10 days or ≥ 30 days"

Types of participants

• Hypertensive encephalopathy and acute renal failure were no longer included as part of the scope of this review. It was assessed that the former had already been addressed in our previous review of hypertensive emergency review. Since acute renal failure is not considered as an acute cardiovascular event according to ICD‐10 (N17‐19), we wanted to avoid ambiguities.

Types of interventions:

• "early treatment" change to  more specific terms "immediate and short‐term treatments"

Types of outcomes:

• From "all cause mortality" changed to more specific terms "all cause mortality at 2, 10 and  ≥ 30 days".

• Non‐fatal serious adverse events were change from a primary outcome to be a secondary outcome due to a critic that there should be only one primary outcome

• Regarding BP and HR changes "during treatment" to more specific terms "during first 24 hours of treatment"

Contributions of authors

Marco Perez formulated the idea for the review, and developed the basis for the protocol. He took the lead roles in searching, identifying and assessing studies, in data extraction and in writing up the review.

Vijaya Musini is an independent reviewer. She helped with the methodology of the review and independently checked the data extraction.

James Wright also acted as an independent reviewer with the extraction and discussion of the included and excluded trials as well as with contributing with the final draft of this review.

Sources of support

Internal sources

• Department of Anesthesiology, Pharmacology & Therapeutics, Universitiy of British Columbia, Canada.

External sources

• CIHR, Canada.

  Salary support

Declarations of interest

None.

Edited (no change to conclusions)