Lee et al reported 2 simple clinical risk scores for predicting severe dengue in an adult soon after onset. These can help to identify a patient who will require more aggressive management. The authors are from Chang Gung University and Kaohsiung Medical University in Taiwan.
Model 1: for dengue illness <= 4 days
Model 2: for dengue illness > 4 days
Parameters:
(1) age in years
(2) WBC count in 10^9/L
(3) minor gastrointestinal hemorrhage
(4) platelet count in 10^9/L
|
Parameter
|
Finding
|
Points Model 1
|
Points Model 2
|
|
age in years
|
< 65 years
|
0
|
0
|
|
|
<= 65 years
|
1
|
1
|
|
WBC count
|
<= 10 * 10^9/L
|
0
|
0
|
|
|
> 10 * 10^9/L
|
3
|
2
|
|
minor GI bleeding
|
no
|
0
|
NA
|
|
|
yes
|
2
|
NA
|
|
platelet count
|
< 100 * 10^9/L
|
0
|
NA
|
|
|
>= 100 * 10^9/L
|
-2
|
NA
|
total score for model 1 =
= SUM(points for all of the parameters)
total score for model 2 =
= SUM(points for all of the parameters)
Interpretation:
• minimum model 1: -2
• maximum model 1: 6
• minimum model 2: 0
• maximum model 2: 3
• A score >= 1 for model 1 had a sensitivity of 0.70 and specificity of 0.91 for severe dengue.
• A score >= 1 for model 2 had a sensitivity of 0.80 and specificity of 0.86 for severe dengue.
Performance:
• The area under the ROC curves for model 1 was 0.85 (0.90 in validation cohort).
• The area under the ROC curve for model 2 was 0.86 (0.92 in validation cohort).
