Description

Atzema et al developed 2 models for predicting 30-day mortality for an adult presenting to the Emergency Department for atrial fibrillation. One model is more complex compared to a simplified model. The authors are from the Institute for Clinical Evaluation Sciences, the University of Toronto, Sunnybrook Health Sciences Centre and University Health Network in Toronto.


Patient selection: adult (range 20 to 105) with atrial fibrillation who is brought to the Emergency Department

 

Outcome: 30-day mortality

 

Parameters:

(1) age (per decade increase)

(2) evidence of bleeding risk (history of gastrointestinal bleeding, hemorrhagic stroke, INR > 3)

(3) history of COPD

(4) smoking history

(5) history of major cancer (any non-skin cancer, melanoma)

(6) history of dementia

(7) chest pain within past 24 hours

(8) systolic blood pressure per 20 mm Hg

(9) heart rate per 10 beats per minute increase

(10) serum troponin result

(11) serum creatinine in µmol/L

(12) other acute ED diagnosis (not primary reason for ED visit; UTI, bronchitis, CKD, etc)

 

Parameter

Finding

Points

age

 

10 per decade increase

bleeding risk

absent

0

 

present

13

history of COPD

absent

0

 

present

11

history of smoking

none

-17

 

former

-9

 

current

0

history of major cancer

absent

0

 

present

15

dementia

absent

0

 

present

15

chest pain

absent

0

 

present

-6

systolic blood pressure

 

-3 per 20 mm Hg increase

heart rate

 

1 per 10 bpm increase

serum troponin

negative

0

 

positive

15

serum creatinine

< 200 µmol/L

0

 

>=200 µmol/L

15

other acute diagnosis

absent

0

 

present

13

 

where:

• Baseline values for age, systolic blood pressure and pulse were not given.

• It seems unlikely that the reference values would be 0.

• Since the simple score gives 1 point for age 65 to 74 and Table 1 shows < 65 as the lowest category, the reference age will be arbitrarily set to 60 in the implementation.

• The mean systolic blood pressure in Table 1 is 134 mm Hg and this will be used as baseline in the implementation.

• The mean pulse in Table 1 was 115 or 117 bpm. 115 will be used as baseline in the implementation.

 

total score =

= SUM(points for all 12 parameters)

 

Interpretation:

• minimum score: negative number, dependent on systolic blood pressure

• maximum score: over 100, dependent on age and heart rate

 

Total Score

Quintile

30-Day Mortality

< 33

1

0.4%

33 to 44

2

0.4%

45 to 54

3

0.2%

55 to 68

4

1.6%

> 68

5

10%

 

Performance:

• The area under the ROC curve is about 0.88.


To read more or access our algorithms and calculators, please log in or register.