Solberg et al developed simple risk scores based on 2,3 and 4 variables that are useful in estimating the 1-year mortality rate following stroke. These can help to identify a patient who may benefit from more aggressive management. The authors are from Akershus University Hospital, The Knowledge Center for the Health Services, Helse Ost Health Services Research Centre and Rikshospitalet University Hospital in Norway.
Patient selection: >= 60 years old with stroke
Outcome: 1-year mortality rate
Parameters:
(1) Canadian Neurological Scale Score for stroke (total score ranges from 0 to 10)
(2) stroke type
(3) Charlson comorbidity score
(4) age in years
|
Parameter |
Finding |
Points |
|
Canadian Neurological Scale |
<= 3.5 |
0 |
|
|
> 3.5 |
1 |
|
stroke type |
ischemic |
0 |
|
|
hemorrhagic |
1 |
|
Charlson comorbidity |
0 |
0 |
|
|
1 |
1 |
|
|
2 |
2 |
|
|
3 |
3 |
|
|
4 |
4 |
|
|
>= 5 |
5 |
2 variable score =
= ((age) – 60) + (30 * (points for CNSS))
3 variable score =
= ((age) – 60) + (30 * (points for CNSS)) + (4 * (points for Charlson comorbidity))
4 variable score =
= ((age) – 60) + (25 * (points for CNSS)) + (5 * (points for Charlson comorbidity)) + (18 * (points for stroke type))
Interpretation:
• minimum score for all 3 scores: 0
• maximum score for 2 variable score: 70 (assumes maximum age is 100 years)
• maximum score for 3 variable score: 90 (assumes maximum age is 100 years)
• maximum score for 4 variable score: 108 (assumes maximum age is 100 years)
• The higher the score the greater the risk of 1-year mortality.
• Results were analyzed based on risk quintiles.
|
Risk Quintile |
2-Variable Score |
Mortality |
|
1 |
0 to 11 |
18% |
|
2 |
12 to 17 |
21% |
|
3 |
18 to 24 |
33% |
|
4 |
25 to 45 |
51% |
|
5 |
46 to 70 |
72% |
|
Risk Quintile |
3-Variable Score |
Mortality |
|
1 |
0 to 16 |
15-21% |
|
2 |
17 to 23 |
16-28% |
|
3 |
24 to 30 |
27-35% |
|
4 |
31 to 50 |
51% |
|
5 |
51 to 90 |
69% |
|
Risk Quintile |
4-Variable Score |
Mortality (derivation) |
|
1 |
0 to 18 |
9-23% |
|
2 |
19 to 25 |
17-31% |
|
3 |
26 to 33 |
34% |
|
4 |
34 to 49 |
48-58% |
|
5 |
50 to 108 |
66-69% |
where:
• The ranges for the 3 and 4 variable models are based on derivation and validation sets.
Performance:
• The area under the ROC curves are in the 0.71 to 0.72 range.
Specialty: Neurology
