Minor haemorrhagic events: There were 2 in each group. In the control group this was 14.1 (95%CI 3.9-116.1) person years per event and in the AMS group 29.9 (95%CI 8.3-246.7) person years per event but the difference is not significant (p=0.451).
Major haemorrhagic events: There were 5 in the control group and non in the AMS group. The control group produced 5.6 (95%CI 2.2-17.4) person years per eventand the AMS group >60 (95%CI 20- ) person years per event, calculated by using an exact test for binomial proportions and computing the probability of observing zero events in the study group under the null hypothesis that the control and study groups have identical underlying event rates. Differences between groups were significant (p= 0.003).
Thromboembolic events: Ten events were recorded in the control group and no events in the AMS group. Person years per event were 2.8 (95%CI 1.5-5.9) in the control group and >60 (95%CI 20- ), calculated as above, in the AMS group (p=<0.0001).
Differences between groups in unplanned clinic visits, emergency room visits and hospital admissions were all very highly significant (p=<0.0001).