Description

Saeed et al developed a score to predict rebleeding after endoscopic treatment of upper gastrointestinal hemorrhage arising from peptic ulcer or other nonvariceal lesion. This can help which patients should be more closely monitored after endoscopy. The authors are from the VA Medical Center and Baylor College of Medicine in Houston, Texas.


 

Parameters used for score:

(1) age of the patient in years

(2) number of concurrent illnesses

(3) severity of concurrent illnesses

(4) site of bleeding at endoscopy

(5) stigma of bleeding at endoscopy

Parameter

Finding

Points

age in years

< 30

0

 

30 – 49

1

 

50 – 59

2

 

60 – 69

3

 

>= 70

5

number of illnesses

0

0

 

1 or 2

1

 

3 or 4

4

 

>= 5

5

severity of illnesses

none

0

 

chronic

4

 

acute

5

site of bleeding

not posterior wall bulb

0

 

posterior wall bulb

4

stigma of bleeding

none

0

 

clot

1

 

visible vessel

3

 

active bleeding (lesion oozing or spurting at endoscopy)

5

after Table 1, page 1843

 

where:

• Number of illnesses: I assume that upper GI hemorrhage is not counted, else there would be no need for a "0" entry.

• Posterior bulb refers to the duodenum.

• The scoring for severity of illness has to be either acute or chronic, since the maximum pre-endoscopy score is 15 (if both could be counted, it would be 19).

• Chronic illness: concurrent chronic life-threatening illness (like cirrhosis)

• Acute illness: presence of concurrent acute-life threatening illness

 

pre-endoscopy score =

= (points for age) + (points for number of illnesses) + (points for illness severity)

 

endoscopy score =

= (points for site of bleeding) + (points for stigma of bleeding)

 

post-endoscopy score =

= (pre-endoscopy score) + (endoscopy score)

 

Interpretation:

• minimum scores: 0

• maximum pre-endoscopy score: 15

• maximum endoscopy score: According to Table 1, the maximum score should be 9. According to Tables 2 and 3, this should be 8.

• maximum post-endoscopy score: 24

• The optimum scores for predicting rebleeding, based on ROC curve analysis (page 1846): pre-endoscopy score >= 8; endoscopy score >= 5; post-endoscopy score >=12.

• To get 100% sensitivity with little loss of specificity, then the pre-endoscopy score >= 6 and post-endoscopy score >= 11 can be used.

 

Score

Cut-Off Value

Likelihood Ratio

Post-Test Probability

pre-endoscopy

>= 9

9.38

55%

 

>= 7

5.50

41%

 

>= 6

3.85

34%

endoscopy

>= 6

0.29

19%

 

>= 4

0.17

14%

post-endoscopy

>= 12

5.50

42%

 

>= 11

4.76

38%

after Table 5, page 1847), with pre-test probability of rebleeding 12%

 


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