Skip navigation

Risk factors for over-anticoagulation with warfarin

 

Clinical bottom line

Patients with an INR of more than 7 were more likely to have a high target INR (3.5) because of prosthetic heart valve, or have had antibiotics prescribed in the previous week.


Reference

S Panneerselvam et al. Analysis of risk factors for over-anticoagulation in patients receiving long-term warfarin. British Journal of Haematology 1998 103: 422-424.

Study

This study was conducted in the Cambridge Anticoagulant Service with decentralised care. It used a computerised decision support system for dose calculations and scheduling INR monitoring. Between mid 1996 and 1997 prospective monitoring identified all INR values above 7, because of increased bleeding risk. Patients and their GPs were questioned about intercurrent illness and change of drug therapy in the four week period before the high INR. Comparison was made with 100 consecutive patients requiring routine INR measurement and warfarin doing.

Results

There were 15,100 INR measurements in the year, and 31 (0.2%) were above 7. The characteristics of the patients with high INR and the consecutive controls are shown in Table 1.

Table 1: patient characteristics for high INR and control

Characteristic

High INR

Control

Odds ratio (95% CI)

Number

31

100

Average age

71

64

Prosthetic heart valve - target INR 3.5

12

8

7.3 (2.6 to 20)

Antibiotic prescribed

5

3

6.2 (1.4 to 28)

Recent drug change

14

17

4.0 (1.7 to 10)

Bleeding in week before test

12

13

4.2 (1.7 to 11)

Major bleeding

5

0

Intercurrent illness

6

6

4.5 (1.3 to 15)

Atrial fibrillation - target INR 2.5

4

39

0.2 (0.1 to 0.7)

The highest risk factor was having a prosthetic heart valve with a target INR of 3.5, but having an antibiotic prescribed, a change in drug therapy, intercurrent illness, and some bleeding in the week before the test were all positively associated. Atrial fibrillation with a target INR of 2.5 was negatively associated with having an INR above 7.

Major bleeding occurred in 5/31 patients with an INR above 7, but none of the 100 controls.

Comment

This is useful stuff, not least because it is good recent UK data garnered from a complete prospective audit of a large service. High INRs above 7 are rare beasts. But to some extent may be preventable because the information from this survey helps to raise awareness of where the risk is. A person with a prosthetic heart valve with some intercurrent illness requiring change of therapy and/or antibiotic is a prime target for some early extra monitoring and dose adjustment.