Lung Injury Prediction Score Is Useful in Predicting Acute Respiratory Distress Syndrome and Mortality in Surgical Critical Care Patients.
Bottom Line: Background.Methods.Conclusion.
Affiliation: Henry Ford Hospital, Detroit, MI 48202, USA.
Background. Lung injury prediction score (LIPS) is valuable for early recognition of ventilated patients at high risk for developing acute respiratory distress syndrome (ARDS). This study analyzes the value of LIPS in predicting ARDS and mortality among ventilated surgical patients. Methods. IRB approved, prospective observational study including all ventilated patients admitted to the surgical intensive care unit at a single tertiary center over 6 months. ARDS was defined using the Berlin criteria. LIPS were calculated for all patients and analyzed. Logistic regression models evaluated the ability of LIPS to predict development of ARDS and mortality. A receiver operator characteristic (ROC) curve demonstrated the optimal LIPS value to statistically predict development of ARDS. Results. 268 ventilated patients were observed; 141 developed ARDS and 127 did not. The average LIPS for patients who developed ARDS was 8.8 ± 2.8 versus 5.4 ± 2.8 for those who did not (p < 0.001). An ROC area under the curve of 0.79 demonstrates LIPS is statistically powerful for predicting ARDS development. Furthermore, for every 1-unit increase in LIPS, the odds of developing ARDS increase by 1.50 (p < 0.001) and odds of ICU mortality increase by 1.22 (p < 0.001). Conclusion. LIPS is reliable for predicting development of ARDS and predicting mortality in critically ill surgical patients.
No MeSH data available.
Related in: MedlinePlus
Mentions: Currently in the literature, a Lung Injury Prediction Score of 4 is considered the cut-off in terms of statistical significance predicting when a patient will develop ARDS . This value, however, is for all ventilated critically ill patients. We were suspicious that the surgical cohort may have a different cut-off LIPS value for predicting a statistically significant risk of developing ARDS. Running the data through a ROC curve and minimizing the difference between sensitivity and specificity, we determined a LIPS value of 7 was statistically significant for predicting ARDS in the surgical critically ill patients (Figure 2). Furthermore, the area under the curve for our ROC was 0.79 (95% confidence interval: 0.74, 0.84), validating LIPS as a strong predictor of ARDS development.
No MeSH data available.