Program Schedule

788
Are Clinical Prediction Scores in Community-Acquired Pneumonia Equally Reliable Across Various Etiologic Categories?

Session: Poster Abstract Session: Clinical Respiratory Infections
Friday, October 10, 2014
Room: The Pennsylvania Convention Center: IDExpo Hall BC
Background: Clinical prediction rules purport to determine, at the time of presentation for community-acquired pneumonia (CAP), which patients (pts) will require ICU admission or die within 30 d. Our goal in the present study was to determine whether the scores are equally reliable in different etiologic subgroups of CAP (i.e. bacterial, viral, undetermined, etc.).

Methods: In this prospective study, we calculated PSI, SMART-COP, IDSA/ATS minor (mATS) criteria, and CURB-65 as predictors of ICU admission and 30 d mortality in pts hospitalized for CAP at the VA Medical Center, Houston. Bacterial CAP (BCAP) was defined as the isolation of a respiratory pathogen from an adequate sputum sample or blood, or urinary antigen tests . Viral CAP (VCAP_ was determined by nasopharyngeal PCR. Pts in whom an infectious etiology was presumed despite negative microbiologic studies were classified as “etiology unknown.”

Results: Of 191 pts with CAP, 23 (12.0%) required ICU admission, 30 d mortality was 7.8% (15/191). Point estimates of sensitivity and area under the receiver-operator-characteristic curve (ROC) as predictors of ICU admission are shown in Tables. The sensitivity of each scoring systems was significantly greater for BCAP compared to CAP of unknown etiology (P<0.05). AUROC for each scoring systems was greater for BCAP than for VCAP and unknown. The same relationship was observed for 30 d mortality.

Conclusion: The sensitivity of each index performed significantly better in pts with BCAP vs. VCAP or CAP of unknown etiology. Sensitivity is more important clinically than AUROC because the consequences of failing to admit to an ICU a pt who requires that level puts the pt at risk for mortality. The high false negative of the scores in viral and CAP of unknown etiology limit their reliability in predicting outcomes.

Etiology

PSI > 3

SMART-COP > 2

mATS > 2

CURB-65 > 2

BCAP

100

89

78

44

VCAP

33

33

67

0

All infected

84

77

69

30

Unknown

50

40

10

20

Sensitivity (%) for Predicting ICU Admission

 

Etiology

PSI

SMART-COP

mATS

CURB-65

BCAP

0.80

(0.65-0.95)

0.88

(0.72-1.00)

0.87

(0.72-1.00)

0.67

(0.47-0.88)

VCAP

0.42

(0.15-0.69)

0.52

(0.08-0.96)

0.58

(0.17-0.99)

0.54

(0.19-0.90)

All infected

0.72

(0.56-0.87)

0.80

(0.64-0.96)

0.79

(0.64-0.85)

0.66

(0.49-0.82)

Unknown

0.50

(0.31-0.69)

0.64

(0.45-0.82)

0.61

(0.43-0.78)

0.58

(0.41-0.75)

AUROC (95% CI) as a Predictor of ICU Admission

Michael S. Abers, BA, Natalie Uy, BA and Daniel Musher, MD, FIDSA, Baylor College of Medicine, Houston, TX

Disclosures:

M. S. Abers, None

N. Uy, None

D. Musher, None

Findings in the abstracts are embargoed until 12:01 a.m. EDT, Oct. 8th with the exception of research findings presented at the IDWeek press conferences.

Sponsoring Societies:

© 2014, idweek.org. All Rights Reserved.

Follow IDWeek