The percentage of patients with major bleeding event.

Major bleeding was defined as any bleed fulfilling one of the following conditions:

• Fatal or life-threatening
• Retroperitoneal, intracranial, intraocular, or intraspinal bleeding (verified by objective testing)
• Bleeding requiring surgical treatment
• Clinically overt bleeding leading to a transfusion (erythrocyte component transfusion or whole blood transfusion) of 4.5 units (equal to 2 units in EU/US) or more
• Clinically overt bleeding leading to a fall in haemoglobin of at least 2 g/dL

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

The percentage of patients with clinically relevant bleeding event.

Any bleed that did not qualify as a major bleed was defined as a minor bleed; minor bleed which fulfilled one of the criteria below was defined as a clinically relevant bleeding event:

• A skin haematoma of at least 25 sqcm
• Spontaneous nose bleed lasting for more than 5 minutes
• Macroscopic haematuria (either spontaneous or, if associated with an intervention, lasting more than 24 hours)
• Spontaneous rectal bleeding (more than spotting on toilet paper)
• Gingival bleeding lasting for more than 5 minutes
• Bleeding leading to hospitalisation
• Bleeding leading to blood transfusion (erythrocyte component transfusion or whole blood transfusion) of less than 4.5 units (equal to 2 units in EU/US)
• Any other bleeding considered clinically relevant by the investigator

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

The percentage of patients with nuisance bleeding event

Any bleed that did not qualify as a major bleed was defined as a minor bleed; all minor bleeding events not fulfilling one of the criteria below was defined as a nuisance bleeding event:

• A skin haematoma of at least 25 sqcm
• Spontaneous nose bleed lasting for more than 5 minutes
• Macroscopic haematuria (either spontaneous or, if associated with an intervention, lasting more than 24 hours)
• Spontaneous rectal bleeding (more than spotting on toilet paper)
• Gingival bleeding lasting for more than 5 minutes
• Bleeding leading to hospitalisation
• Bleeding leading to blood transfusion (erythrocyte component transfusion or whole blood transfusion) of less than 4.5 units (equal to 2 units in EU/US)
• Any other bleeding considered clinically relevant by the investigator

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

Percentage of patients with the composite clinical endpoint (ischemic or haemorrhagic stroke (fatal or non-fatal), transient ischemic attacks, systemic embolism, myocardial infarction (fatal or non-fatal), other major adverse cardiac events, and death)

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

The percentage of patients with ischemic or haemorrhagic stroke (fatal or non-fatal)

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

The percentage of patients with transient ischemic attack

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

The percentage of patients with systemic embolism

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

The percentage of patients with myocardial infarction (fatal or non-fatal)

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

The percentage of patients with other major adverse cardiac events

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

The percentage of patients with death

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

Intensity of event is categorised as mild, moderate and severe.

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

Discontinuation of the study drug due to adverse events.

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.

The blood coagulation parameter aPTT was assessed in patients allocated to the dabigatran etexilate groups at week 0, prior to drug administration and at the trough at week 1, 4 and 12.

Full Analysis Set

The blood coagulation parameter ECT was assessed in patients allocated to the dabigatran etexilate groups at week 0, prior to drug administration and at the trough at week 1, 4 and 12.

Full Analysis Set

The blood coagulation parameter INR was assessed in patients allocated to the dabigatran etexilate groups at week 0, prior to drug administration and at the trough at week 1, 4 and 12.

Full Analysis Set

Analysis based on concomitant use of aspirin compared to no aspirin users. 11-dehydrothromboxane B2 is measured in urine of patients.

Per Protocol Analysis Set

Full Analysis Set

The number of patients with ALT, AST, alkaline phosphatase, or bilirubin exceeded the upper limit of normal (ULN) range

Safety set was used for safety endpoints. Full analysis set was used for efficacy endpoints. The safety set comprises all patients who were treated with trial medication at least once. The full analysis set comprises all patients who were randomised and treated with trial medication at least once. No data was imputed.