APSI 1997.

Study characteristics
Methods	Randomised multi‐centre trial at 44 clinical centres in France between April 1987 and September 1988
Participants	607 patients met the following inclusion criteria: (1) had > 2 of the 3 classic signs of AMI, that is, typical chest pain of 11 hours in duration, typical Q waves, and significant release of cardiac enzyme(s); (2) were admitted for this acute event > 2 and < 22 days before; and presented with > 6 of the secondary risk factors of the selection algorithm, including “major” secondary risk factor (i.e. before documented AMI, history of dyspnoea when walking on flat ground, documented atrial fibrillation, ventricular fibrillation, ventricular tachycardia, overt heart failure or sinusal tachycardia during the reference event, recurrent AMI or angina pectoris before the eighth day) Male:female = 443:164 Mean age = 62.9 years Exclusion criteria: heart rate < 45 beats/min; complete auriculoventricular block and acute heart failure that required treatment with > 2 drugs of different classes (e.g. diuretics, vasodilators) (if the condition disappeared before the twenty‐second day, the patient could be included); contraindication to beta‐blocking treatment; non‐cardiac disease with poor prognosis; impossibility to participate; indication for beta‐blocking treatment; age > 75 years; death; malignancy; valvular disease; coma; asthma; chronic broncho‐pneumopathy; Raynaud syndrome; participation in another study; enrolled in APSI before
Interventions	Experimental group: acebutolol (200 mg twice daily for 12 months) (n = 298) Control group: placebo (n = 309) Co‐intervention: 30% of patients had taken aspirin regularly; one‐half were given oral anticoagulants; one‐third were given diuretics, whereas 10% received digitalis. Finally, 39.3% in the acebutolol group and 37.9% in the placebo group received nifedipine Excluded medication: not described
Outcomes	Primary: total death (reported at a median of 5 years of follow‐up) Secondary: cardiovascular death (sudden death, fatal + non‐fatal myocardial reinfarction, cardiac failure, stroke or cerebral haemorrhage) (reported at 12 months' follow‐up) Time points reported: median of 5 years and 12 months
Notes	Study author was contacted on 10‐02‐2017 at jpb@upcl.univ‐lyon1.fr and jean‐pierre.boissel@novadiscovery.com. No response was received Mean time from onset of symptoms to randomisation was 10.5 days In the first publication from 1990, total mortality in the placebo group is 34, and in the second publication from 1997, the same mortality at 1 year (37) was reported in Table 1. However, because 34 deaths were reported in 2 out of 3 publications, this is the number we used MACE was calculated and was reported as a composite of 'non‐fatal reinfarction + cardiovascular mortality' Study was funded by a grant from SPECIA Pharmaceuticals, Paris, France. However, it is said that "the Policy Board monitored the trial results on an ongoing basis. Its voting members were not otherwise involved in the trial and had no connection with the sponsor". Furthermore, the study was initiated by the French Society of Cardiology
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Central randomisation was achieved with a Minitel@, a terminal linked to the telephone system allowing an on‐line remote data entry"
Allocation concealment (selection bias)	Low risk	"After editing of the eligibility criteria by the central computer, the treatment number was shown on the screen. Random permutations with blocks of 6 were used for randomisation. Treatment numbers were not sequential"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"Tablets of acebutolol and placebo were indistinguishable. A round‐the‐clock telephone service was available for investigators who needed to know the treatment. The code could only be broken in case of life‐threatening conditions, the care of which required to know the treatment given to the patient"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"The coordinating and data handling center was in charge of managing the trial, editing collected data, controlling compliance by both investigators and patients, randomising the treatment through an on‐line computerized procedure and analysing the results" Nobody but the voting members of this committee and 4 individuals in the coordinating centre were aware of the interim results. The critical events committee met periodically to classify the causes of death. Its members were not otherwise involved in the trial and were kept blind vis‐a‐vis study treatment
Incomplete outcome data (attrition bias) All outcomes	High risk	None of the 607 included and randomised patients were lost to follow‐up at 12 months' follow‐up; however, at maximum follow‐up, which is our primary time point of interest, 15/298 (5.0%) from the acebutolol group and 6/309 (1.9%) from the placebo group were lost to follow‐up
Selective reporting (reporting bias)	Unclear risk	No protocol was found; however mortality was reported
Other bias	Low risk	No other biases were found