Table 1. Details of randomised trials.

Source		Dose of aspirin and duration of follow-up (range, mean or median)	Number of subjects		Number of bleeds		Fatal bleeds
			Aspirin	No aspirin	Aspirin	No aspirin	Aspirin	No aspirin
Peto et al (1988)[34]	5,139 healthy male doctors	500mg daily or 300 enteric coated for 5–6 years	3,429	1,710	89	27	3	3
Physicians’ Health Study (1989)[35]	22,071 healthy physicians (United States)	325 mg alternate days for 60.2 months	11,037	11,034	13	6	1	0
Swedish Angina Trial (1992)[36]	1,360 patients with a transient ischaemic attack or myocardial infarction	75 mg daily for 32 months	676	684	9	4	1	1
Internat. Stroke Trial (1997)[37]	19,435 Patients with ischaemic stroke	300 mg daily for 6 months	4,858	4,860	23	14	4	2
Thrombosis Prevention trial (1998)[38]	5499 men at increased risk of vascular disease	75 mg daily for 6.8 years	1,268	1,272	7	4	1	2
Hansson et al (1998)[30]	Hypertensive patients on ‘optimal’ treatment	75 mg daily for 3.8 years	9,399	9,391	129	70	5	5
Primary Prev. Project (2001)[40]	4,495 selected from general practitioner lists	100 mg daily for 3–6 years	2,226	2,269	17	5	1	3
Baron et al (2003)[41]	1,121 patients selected at colo-rectal screening	81 mg daily 325 mg daily each for 3 years	377/372	372	2/4	3	0/1	0/0
Ridker (2005)[42]	39,876 women	100 mg alternate days for 10 years	19,934	19,942	127	91	2	3
Belch et al (2008)[43]	1,276 diabetic patients with arterial disease	1900 mg daily for 6.7 years	638	638	28	31	0	2
Brighton et al (2012)[44]	822 patients with venous thrombosis	100 mg daily for 37 months	411	411	8	6	0	2
Risk of a GI bleed
	On aspirin	8/1000	RR = 1.55 (1.32, 1.83) heterogeneity p = 0.401
	On placebo	5/1000
Risk of a bleed being fatal
	On aspirin	4%	RR = 0.45 (0.25, 0,80) heterogeneity p = 0.783
	On placebo	10%
Risk of a fatal bleed if randomised to aspirin
	On aspirin	3.7/10,000	RR = 0.77 (0.41, 1.43) heterogeneity p = 0.908
	On placebo	4.7/10,000

Abbreviations: CI: confidence interval; mg: milligrams; RCT: randomised controlled trial; RR: risk ratio.