. 2012 Jan 23;141(2 Suppl):e419S–e494S. doi: 10.1378/chest.11-2301

Table 19.

—[Section 3.1.1-3.1.4] Estimated Absolute Difference in Recurrent VTE and Major Bleeding Events (Including Fatal Events) With 5 Years of vs No Extended Anticoagulation

	Outcomes After 5 y of Treatment	Risk of Bleeding
	Outcomes After 5 y of Treatment	Low	Intermediate	High^a
First VTE provoked by surgery	Recurrent VTE reduction per 1,000	↓ 26 (19-27) (1 fatal)^b	↓ 26 (19-27) (1 fatal)^b	↓ 26 (19-27) (1 fatal)^b
First VTE provoked by surgery	Major bleeding increase per 1,000	↑ 24 (2-73) (3 fatal)^b	↑ 49 (1-173) (5 fatal)^b	↑ 98 (1-346) (11 fatal)^b
First VTE provoked by a nonsurgical factor/first unprovoked distal DVT	Recurrent VTE reduction per 1,000	↓ 132 (93-137) (5 fatal)^c	↓ 132 (93-137) (5 fatal)^c	↓ 132 (93-137) (5 fatal)^b
	Major bleeding increase per 1,000	↑ 24 (2-73) (3 fatal)^c	↑ 49 (1-173) (5 fatal)^c	↑ 98 (1-346) (11 fatal)^b
First unprovoked proximal DVT or PE	Recurrent VTE reduction per 1,000	↓ 264 (186-273) (10 fatal)^d	↓ 264 (186-273) (10 fatal)^d	↓ 264(186-273) (10 fatal)^b
First unprovoked proximal DVT or PE	Major bleeding increase per 1,000	↑ 24 (2-73) (3 fatal)^d	↑ 49 (1-173) (5 fatal)^d	↑ 98 (1-346) (11 fatal)^b
second unprovoked VTE	Recurrent VTE reduction per 1,000	↓ 396 (279-409) (14 fatal)^e	↓ 396 (279-409) (14 fatal)^d	↓ 396 (279-409) (14 fatal)^c
second unprovoked VTE	Major bleeding increase per 1,000	↑ 24 (2-73) (3 fatal)^e	↑ 49 (1-173) (5 fatal)^d	↑ 98 (1-346) (11 fatal)^c

Recommendations:

Risk of dying in patients with a recurrent VTE or a major bleed:

• Case fatality rate of recurrent VTE after discontinuing oral anticoagulation therapy: 3.6% (Carrier et al¹²).

• Case fatality rate of major bleeding during initial oral anticoagulation therapy: 11.3% (Carrier et al¹²) (no data available for after discontinuing oral anticoagulation therapy).

Annual risks of recurrent VTE after discontinuation of anticoagulation:

• First VTE provoked by surgery: 1% (Iorio et al¹⁷¹); we assumed a 0.5% yearly risk thereafter (3% over 5 y).

• First episode of VTE provoked by nonsurgical factor: ∼5% the first year (Iorio et al¹⁷¹); we assumed a 2.5% yearly thereafter (15% over 5 y)

• First episode of unprovoked VTE: 9.3% over 1 y (Rodger et al¹⁸⁵); 11.0% over 1 y, 19.6% over 3 y, 29.1% over 5 y (Prandoni et al²⁰⁸). We assumed a risk of 10% the first year after discontinuation and 5% yearly thereafter (30% over 5 y).

• Second episode of unprovoked VTE: we assumed that this inflicts 1.5 the risk of recurrent VTE relative to first episode of unprovoked VTE: 15% the first year after discontinuation, 7.5% yearly thereafter (45% over 5 y).

Relative risk reduction with extended anticoagulant therapy:

• 82% based on Table 18

Annual risks of major bleeding in patients not on anticoagulant therapy:

• Low risk, 0.3%/y; intermediate risk 0.6%/y; high risk, 2.4%/y (Table 2).

Relative risk of major bleeding with extended anticoagulant therapy:

• 2.6 based on Table 18.

Criteria used to decide on direction and strengths of recommendations:

• Criterion for a strong recommendation against whenever the estimated number of fatal bleeding events exceeded the estimated number of fatal recurrent VTE prevented.

• Criterion to go from a strong recommendation against to weak recommendation against: difference between the lower boundary of increased major bleeding and upper boundary of reduction in recurrent VTE < 2% (risk over 5 y averaged per year).

• Criterion to go from a weak recommendation against to a weak recommendation in favor of: difference between point estimate of reduction of recurrent VTE and point estimate for increase in major bleeding is > 2% (risk over 5 y averaged per year) (2% to account for the burden and cost of VKA).

• Criterion to go from a weak recommendation for to strong recommendation for: difference between the lower boundary of reduction in VTE and upper boundary of increased major bleeding > 4% (risk over 5 y averaged per year).

Another way of interpreting the direction and strength of recommendation based on the number of deaths (related to either bleeding or recurrent VTE) is as follows:

• A strong recommendation against: extended anticoagulation is estimated to be associated with an increase in deaths.

• A weak recommendation against: extended anticoagulation is estimated to be associated with from no effect on deaths to only a very small reduction in deaths (0-4/1,000 prevented over 5 y or < 0.5%/patient-y).

• A weak recommendation for: extended anticoagulation is estimated to be associated with a small reduction in deaths (5 to 9/1,000 prevented over 5 y or 0.5%-0.9%/patient-y).

• A strong recommendation for: extended anticoagulation is estimated to be associated with a large reduction in deaths (> 10/1,000 prevented over 5 y or > 1%/patient-y).

With an eightfold risk of bleeding in the high-risk group compared with the low-risk group, a strong recommendation against extended anticoagulation for a second unprovoked VTE is justified. The high-risk group, however, includes patients who have a risk of bleeding that is less than this estimate (eg, patients aged > 75 y without additional risk factors for bleeding [Table 2]) and, therefore, may benefit from extended anticoagulant therapy. For this reason, we provide a weak rather than a strong recommendation against extended anticoagulation for patients with a second unprovoked VTE in the high-bleeding-risk group.

Strong against

Weak against

Weak in favor

Strong in favor