Evaluation of severity scoring models in mortality prediction of severe community acquired pneumonia.
Date Issued
2023-09
Author(s)
Rangelov Goran
Abstract
Background: community-acquired pneumonia (CAP) is one of the most common infection worldwide and major cause of death particularly in elderly and patients with comorbidities. Severe CAP (sCAP) refers to acute lung infection that require intensive care treatment due to disease severity. Recently, two major criteria are widely accepted to define sCAP: need of mechanical ventilation or need of vasopressors for shock. Several prediction models are designed for assessment of disease severity and for early prediction of pneumonia outcome.
Aim of this study was to evaluate the mortality prediction capacities of the most commonly used severity scoring models in patients with sCAP.
Material and methods: the study included 129 adult patients with sCAP hospitalized at ICU within Clinic for Infectious Diseases in Skopje during a 3-year period. Primary outcome was 30-day in-hospital mortality. Demographic, clinical and biochemical parameters were recorded and seven severity scores were calculated: Charlson comorbidity Index, CURB 65, SCAP at admission, SAPS II and APACHE II after 24 hours, MPM and SOFA during the first 48 hours. Receiver Operating Curve (ROC) analysis was performed and areas under the curve (AUC) were calculated to evaluate the prediction capacities of analysed scores.
Results: the mean age of studied patients was 61 year and 66,7% were males. An overall mortality was 43.4%. Charlson Comorbidity Index has higher value in non-survivors and it was associated with poor outcome. All scores showed significantly increased values among non-survivors, except CURB-65 that had similar results in both groups, and no significance with the outcome. AUC for all scores had close values, ranging from 0.714 for APACHE II to 0.755 for SCAP. The highest AUC showed MPM and SOFA when calculated at 48 hour upon admission, with values of 0.800 and 0.839 respectively.
Conclusion: the results of our study showed that the most commonly used severity scoring models had great ability to identify the patients with pneumonia that had increased risk for poor outcome, however, none of them presented stronger capacity over the others to predict the disease mortality.
Aim of this study was to evaluate the mortality prediction capacities of the most commonly used severity scoring models in patients with sCAP.
Material and methods: the study included 129 adult patients with sCAP hospitalized at ICU within Clinic for Infectious Diseases in Skopje during a 3-year period. Primary outcome was 30-day in-hospital mortality. Demographic, clinical and biochemical parameters were recorded and seven severity scores were calculated: Charlson comorbidity Index, CURB 65, SCAP at admission, SAPS II and APACHE II after 24 hours, MPM and SOFA during the first 48 hours. Receiver Operating Curve (ROC) analysis was performed and areas under the curve (AUC) were calculated to evaluate the prediction capacities of analysed scores.
Results: the mean age of studied patients was 61 year and 66,7% were males. An overall mortality was 43.4%. Charlson Comorbidity Index has higher value in non-survivors and it was associated with poor outcome. All scores showed significantly increased values among non-survivors, except CURB-65 that had similar results in both groups, and no significance with the outcome. AUC for all scores had close values, ranging from 0.714 for APACHE II to 0.755 for SCAP. The highest AUC showed MPM and SOFA when calculated at 48 hour upon admission, with values of 0.800 and 0.839 respectively.
Conclusion: the results of our study showed that the most commonly used severity scoring models had great ability to identify the patients with pneumonia that had increased risk for poor outcome, however, none of them presented stronger capacity over the others to predict the disease mortality.
File(s)![Thumbnail Image]()
![Thumbnail Image]()
Loading...
Name
20231024_234826.jpg
Size
2.62 MB
Format
JPEG
Checksum
(MD5):4c6af79957bc18aee7e1a260af69d3a8
Loading...
Name
20231024_234614.jpg
Size
1.02 MB
Format
JPEG
Checksum
(MD5):df3f3dc2d3eec0013d633eec01b2f211