Skip navigation

Anticoagulation works in clinical practice

Clinical bottom line

Anticoagulation with warfarin in clinical practice confers the same reduction in strokes found in randomised trials.


Reference

JJ Caro et al. Anticoagulant prophylaxis against stroke in atrial fibrillation: effectiveness in clinical practice. Canadian Medical Association Journal 1999 161: 493-497.

Study

This prospective study examined all patients undergoing electrocardiography between 1990 and 1993 in two hospitals in Montreal. There was a consistent diagnosis for atrial fibrillation. Excluded were patients with mitral stenosis, artificial valves or heart transplant. Other reasons for exclusion were refusal to participate, severe illness, transient atrial fibrillation, language barrier, inability to be followed up or death.

Follow up ECG and telephone interview was carried out every six months. Risk factors for stroke were collected, plus any information about stroke or transient ischaemic attacks (TIA), or major bleeding events. The latter were defined as fatal bleeds or those requiring packing or transfusion. Patient reported information was supplemented with chart review. Follow up ended in June 1994, giving a maximum follow up fo five years.

Results

There were 1725 patients with atrial fibrillation originally screened, of whom 221 were eligible for follow up. Average follow up was 27 months, with a total of 500 patient years of data. The mean age of patients was 72 years. Treatment was none in 38, warfarin in 87, aspirin in 31 and blended in 65. Blended treatment was used when patients switched treatments during follow up.

There were 19 strokes and six TIAs during follow up. Patients receiving warfarin had a significantly lower risk of stroke or TIA (or combined stroke plus TIA) than no treatment, aspirin or blended treatment (Figure 1). The relative risk of stroke compared with no treatment was 0.31 (95%CI 0.09 top 1.00), and for stroke plus TIA it was 0.34 (0.12 to 0.99).

The rate for any bleeding event was higher for warfarin (Figure 1), and was significantly higher than no treatment, with a relative risk of 1.8 (95% CI 1.1 to 3.0). Four patients with warfarin had a major bleed, compared with one without treatment.

The combined rates of stroke plus TIA plus major bleed is shown in Figure 2, and the total number of strokes, TIA and major bleeds in Table 1.

Figure 1: Strokes, strokes plus TIA and any bleeding event


Figure 2: Combined stroke plus TIA plus any bleeding event


Table 1: Stroke, TIA and major bleeds

Event

No treatment

Warfarin

Aspirin

Blended

Number of patients

38

87

31

65

Stroke

7

4

4

4

TIA

1

1

0

3

Major bleed

1

4

0

0

Crude total event rate

8/38 (21%)

9/87 (10%)

4/31 (13%)

7/65 (9%)

 

Comment

This study had the strength of being prospective and with regular follow up, but was limited by its size. It broadly supports the results of randomised trials, with warfarin reducing the number of strokes, but at a cost of more bleeding events. A key aspect of treatment decision is likely to be patient attitudes to the relative benefits and disbenefits of strokes and bleeds. Patients appear to rate strokes as being much worse than bleeds.