Table 1.

Randomised trials of oral anticoagulants versus antiplatelet treatments in non-rheumatic atrial fibrillation included in systematic review

Trial	Patient characteristics	Intervention (months of follow up)	INR	Random allocation	Blinding used	Withdrawals from trial (warfarin/antiplatelet)
AFASAK 115	Chronic non-rheumatic atrial fibrillation, no stroke in prevoius month but 5.5% had previous TIA/stroke, BP <180/100 mm Hg; n=671, mean (range) age 73 (38-91) years	Aspirin 75 mg v warfarin (24)	2.8-4.2	Drug packages sequenced randomly by computer; concealment of allocation may be inadequate	Outcomes not assessed or reviewed blind to treatment; statistician blinded	38% (126) v 13% (44)
SPAF II <7517	Non-rheumatic atrial fibrillation ⩾12 months. No TIA/stroke in previous 2 years but 7% had previous TIA/stroke, n=715, all patients <75, mean (SD) 64 (8) years	Aspirin 325 mg v warfarin (36)	2-4.5	Computerised randomisation performed in each centre; concealment adequate	Outcomes reviewed blind; not stated for statistician	Withdrawals not reported but only 0.8 % (3) v 0.3% (1) lost to follow up (<0.4% overall)
SPAF II 75+17	As SPAF II <75 but all aged ⩾75 years, n=385, mean (SD) 80 (3) years	Aspirin 325 mg v warfarin (24)	2.4-4.5	Computerised randomisation performed in each centre; concealment adequate	Outcomes reviewed blind; not stated for statistician	Withdrawals not reported; no losses to follow up
AFASAK 216	Non-rheumatic atrial fibrillation, third <1 year duration, 70% diagnosed heart failure, 8% past TIA/stroke, 8% history of MI, n=239, mean (SD) age 73 (7) years	Aspirin 300 mg v warfarin (42)	2.0-3.0	Computerised randomisation performed; concealment adequate	Outcomes reviewed blind; not stated for statistician	Overall 16.5% withdrawn and 8.6% dropped out, but not broken down by treatment allocation; losses to follow up not reported
SIFA18	Non-rheumatic atrial fibrillation or paroxysmal AF, 50% had previous TIA/stroke, n=916, mean (SD) age 73 (8) years	Indoprofen 400 mg v warfarin (12)	2-3.5	Computerised dial up telephone randomisation service; concealment adequate	Outcomes reviewed blind; not stated for statistician	10.6% (48) v 9.1% (42)
PATAF19	Non-rheumatic atrial fibrillation or paroxysmal AF drawn from primary care, 9% had history of MI, n=272, mean (SD) age 75 (7.5) years	Aspirin 150 mg v warfarin (31)	2.5-3.5	Computerised dial up telephone randomisation service; concealment adequate	Outcomes reviewed blind; not stated for statistician	No losses to follow up

INR=international normalised ratio, MI=myocardial infarction, TIA=transient ischaemic attack, AF=atrial fibrillation.