Barnett-Griness et al reported a model for predicting major bleeding in a patient being treated with a non-warfarin oral anticoagulant. The authors are from Lady Davis Carmel Medical Center and Technion Israel Institute of Technology in Haifa, Israel.
Patient selection: non-warfarin (new oral anticoagulant or NOAC) for a patient with atrial fibrillation
Parameters:
(1) sex
(2) hemoglobin in g/dL
(3) platelet count per µL
(4) concurrent antiplatelet therapy
(5) hypertension
(6) prior major bleeding
(7) known fall risk
(8) serum cholesterol in mg/dL
(9) eGFR in mL/min
Parameter
Finding
Points
sex
female
0
male
7
hemoglobin
>= 13 and male
0
< 13 and male
6
>= 12 and female
0
< 12 and female
5
platelet count
>= 100,000 per microliter
0
< 100,000 per microliter
9
concurrent antiplatelet rx
no
0
yes
7
hypertension
no
0
yes
8
history major bleeding
no
0
yes
5
known fall risk
no
0
yes
4
serum cholesterol
>= 240
0
200 to 239
1
160 to 199
2
< 160
5
eGFR
>= 90
0
60 to 89
4
30 to 59
6
<= 29
8
total score =
= SUM(points for all of the parameters)
Interpretation:
• minimum score: 0
• maximum score: 59
• The higher the score the greater the risk of bleeding.
Score
Predicted Bleeding
0 to 9
0.3 to 0.5%
10 to 20
0.6 to 1.0%
21 to 26
1.1 to 1.5%
27 to 31
1.6 to 2.0%
32 to 38
2.1 to 3.0%
39 to 43
3.1 to 4.0%
44 to 46
4.3 to 4.8%
47 to 49
5.1 to 5.8%
50 to 52
6.1 to 6.9%
53 or 54
7.3 to 7.7%
55 or 56
8.2 or 8.7%
57 or 58
9.2 or 9.7%
59
10.3%
Performance:
• The area under the ROC curve ranged from 0.62 to 0.65.
To read more or access our algorithms and calculators, please log in or register.
