Trillo-Alvarez et al reported an acute lung injury prediction score (LIPS). This can help to identify a patient at risk for acute lung injury. The authors are from the Mayo Clinic in Rochester.
Predisposing conditions:
(1) sepsis
(2) shock
(3) trauma
(4) pneumonia
(5) aspiration
(6) pancreatitis
(7) surgery
Risk modifiers:
(1) alcohol abuse
(2) smoking
(3) hypoalbuminemia
(4) diabetes mellitus
(5) chemotherapy
(6) FIO2
(7) tachypnea (> 30 breaths per minute)
|
Parameter |
Finding |
Points |
|
sepsis |
absent |
0 |
|
|
present |
1.5 |
|
shock |
absent |
0 |
|
|
present |
1.5 |
|
trauma |
absent |
0 |
|
|
present |
0.5 |
|
pneumonia |
absent |
0 |
|
|
present |
0.5 |
|
aspiration |
absent |
0 |
|
|
present |
1.5 |
|
pancreatitis |
absent |
0 |
|
|
present |
1.5 |
|
surgery |
none |
0 |
|
|
elective |
1.5 |
|
|
emergency |
2.0 |
|
alcohol abuse |
absent |
0 |
|
|
present |
1.0 |
|
smoking |
absent |
0 |
|
|
present |
0.5 |
|
hypoalbuminemia |
absent |
0 |
|
|
present |
2.0 |
|
diabetes mellitus |
absent |
0 |
|
|
present |
-1.5 |
|
chemotherapy |
no |
0 |
|
|
yes |
2.0 |
|
FIO2 |
0.21 to 0.35 |
0 |
|
|
> 0.35 |
1.0 |
|
tachypnea |
absent |
0 |
|
|
present |
1.0 |
where:
• Several combinations require assumptions. For example pneumonia and aspiration are assumed to be separate.
• An oxygen flow > 4 liters per minute by nasal cannula is considered an FIO2 > 0.35.
• The cutoff for hypoalbuminemia is not stated but presumed to be < 3.5 based on later papers.
total score =
= SUM(point for all of the parameters)
Interpretation:
• minimum score: -1.5
• maximum score: 16.5
• A score > 3 was considered predictive of acute lung injury.
Performance:
• The area under the ROC curve is claimed to be 0.84 but the sensitivity is 0.41 and specificity 0.90 in the derivation set (sensitivity 0.69 and specificity 0.84 in validation cohort).
Specialty: Pulmonology
